I am putting these designs up in the forum prior to placing them in the Projects section to get feedback.
Attached are circuit designs for a 4 channel PWMdew heater controller. This circuit basic design has been around for some time, however, principally each previous design has a major flaw in that these circuits utilise a low side switch device. This design means that the outside shell of your RCA connector can effectively become +12V. Touch this against any earth (which is easy to do with RCA connectors) and you will pass 12V thru to earth with possible consequences to your valuable equipment.
My circuit utilizes a “P” channel FET acting as a “high side switch”. In this manner, the outside of the RCA connector will always be earth. You can not short the +12V because in my design the outside of the RCA connector will always be 0V as it connects the RCA jack outer sheath to earth.
I have provided this project in two circuit PCB versions. Both are 4 independent channel PWMdew heaters: one fits on a single side 100mm x 60mm PCB; the other has fuse protection for each FET and fits on a single side 100mm x 80mm PCB. This allows other users to amend my circuit easily, print and make their own PCBs in the future if they so wish. These PCB designs are within the Eagle freeware maximum PCB size parameters of 100mm x 80mm.
I have made some small changes to my previous design attempts and have trialled many other changes but settled on: a simple design; reliable design; cheap with easily available parts; that fit on a PCB using Eagle PCB freeware! As well the design is a 4 independent channel PWM heater. You can control how much energy you apply to keep dew at bay at independent levels for each individual channel.
edit:
Oh the drawing is showing a dual channel only just to show the design.
Eagle free-ware is available here
Last edited by wasyoungonce; 18-07-2010 at 01:19 PM.
Part 2:
I envisage that I will develop this circuit further possibly incorporating current fold back and other changes. When I can get my hands on some of the newer protected power switch FETs. But at the moment these designs are released for use.
I have done many hrs of testing to ensure the circuits are capable.
Circuit Specifications (an overview):
Will operate between 10V and 18V DC. No over voltage protection. It’s not needed here, ATM.
4 Independent channels with control of pulse width mark space ratios from 1% to 99%.
Can deliver 3 Amps (plus) per channel but recommend all 4 channels in operation do not exceed 8 amps total. You have to remember this is a cheap design and 3 amps is around 36 Watts of energy. That is a lot! 99.9% of users will never use anywhere near this amount of energy to keep dew at bay!
FETs run at around 55ºC (case temp) at max I draw of 3A (with Heat-sink).
Over current protection (4 amp fuses) on the 100mm x 80mm PCB design for each individual FET.
Pulse width (energy output settings) can be easily identified by the use of a flashing LED (inbuilt in each switch) on each channel.
Low frequency PWM. Around 1Hz of operation which should not induce “spike” power supply noise. IMHO you do not need higher operating frequencies here.
Low power consumption. Around 4 mA at idle.
Uses a “P” channel high side switch with low “on” resistance (Rds). You cannot short +12V. The FET, because of its low Rds, will only be dissipating a few milli-watts of energy over its package, meaning the majority of the energy is delivered across the heater strap thus maximum efficiency.
Cheap design. Can be made for around $100 or less depending upon your access to odds & ends.
Attached are pics of my 1st version PCB & design. These pics are just to show how I fitted the PCB. A lot has changed/improved since this version however; it is still running well & without fault for over 1 year.
When all is said & done I’ll fill in the ground planes on the PCBs and transport them to the projects & Articles.
edit:
the box design is for the small PCB only. I have never built the large PCB...but will do so soon.
Last edited by wasyoungonce; 18-07-2010 at 01:21 PM.
It may be useful to someone or it may provide ideas to someone else.
In any case the design has it's pros & cons but principally it's cheap and easy to understand & make.
I am looking at a 6 channel version as the cct is easily scalable to a bigger version. This would suit even the most ardent user. But, I get the impression that if I go this far then maybe AVR controlled (like hikerbob's Arduino) would be a better idea.
I also meant to ask..if you have suggestions cct change ideas..or such like please feel free to speak up.
Last edited by wasyoungonce; 18-07-2010 at 02:36 PM.
Just two comments
Re low side switching and shorting, if you use "unswitched" RCA connectors on the controller, you can always wire the 12V to the centre, and this makes it much more difficult to accidentally short out
I have attached a simple circuit i use based on an LM324 opamp
( as well as a finished unit to show its not complicated )
Three sections of the OPAmp make up a proper sawtooth output waveform, and the 4th section is used to provide a buffered power supply to drive the Mosfet.
By adding one more LM324, you can use the sawtooth waveform to feed 4 more supplies and hence get 5 independent channels out.
I would have to say tho, that after playing with multichannel devices like this, the cabling becomes a nightmare
What i am looking at doing is to make a very small PCB using SMT components, and actually build a controller into each dewstrap.
That way, all that is reqd is a simple 12V supply bus.
Just two comments
Re low side switching and shorting, if you use "unswitched" RCA connectors on the controller, you can always wire the 12V to the centre, and this makes it much more difficult to accidentally short out
..
Andrew
Andrew...I have checked this and any PWM low side switch...will see a voltage on the outside of the RCA shell when you lower thePWM output.
It doesn't matter if you wire the Vcc to the inner core...the outer shield will become Vcc with the heater output device off or at very low mark space ratios.
What i am looking at doing is to make a very small PCB using SMT components, and actually build a controller into each dewstrap.
That way, all that is reqd is a simple 12V supply bus.
will see a voltage on the outside of the RCA shell when you lower thePWM output.
I agree that if a dewstrap is fitted, the outer shell can become live when the unit turns off, but its also much harder to short it.
I have actually seen low side units where the outer shell ( in the controller ) was wired to 12V , hence was always live, even when not connected.
I only use proper 2.5mm DC connectors on my units, just to avoid this.
I agree that if a dewstrap is fitted, the outer shell can become live when the unit turns off, but its also much harder to short it.
I have actually seen low side units where the outer shell ( in the controller ) was wired to 12V , hence was always live, even when not connected.
I only use proper 2.5mm DC connectors on my units, just to avoid this.
Andrew
I agree that RCA connectors are....at best...indeed a bad choice> However I kept the design within was was currently available for comercial straps and indeed most amateur equipment.
An easy fix is to heatshrink cover your RCA jacks & connectors but I know only too well some day some one will be very unhappy due to a short... I have practised Murphy's law more than once.
The solution is a high side switch. The problem is getting a low rds on FET ("P channel")to handle this current.
wasyoungonc , I've not opened up the files yet but thanks for posting this. I want to look and see what I can learn.
AndrewJ, I'd pondered options for putting the switching components at the heater strap. I did some investigation of other 1-wire devices to see if there was a neat way to controlPWM from one while using a central controller (not found it yet).
In the end I decided it was complicating things and giving little benefit for what I was trying to achieve. In hindsight I probably could have put an ATTiny45 (or similar) at each heater strap with a voltage regulator, a temp sensor and fet in a fairly small package. Add a pot to the mix and you could have a compact fully contained dew controller in each strap. I've not explored that yet but for an idea of what can be done with 8pin Micro this looked interesting http://www.circuitlake.com/usb-wirel...ty-sensor.html . There might even still be a line or two left to communicate back to a central controller (humidity sensing, display etc) if required. Lot's of options left to learn and play.
I did some investigation of other 1-wire devices to see if there was a neat way to controlPWM from one while using a central controller (not found it yet).
I was looking at using a simple SMT pic/picaxe or the like to put the controller on the strap. By using 2 DS18B20 one wire sensors to get ambient and strap temps, and one PWM output to drive the FET, a pretty simple and small footprint "thermostatic" design can be made.
I have made several controllers like this ( including humidity sensing ), but they were all with big components and required big boxes.
A 4 channel PCB design & 6 Channel PCB design. I have included a parts list and circuit description document. All documents are zipped for smaller size.
I'll be building a 6 channel PCB soon as I need the extra ccts.
I'll get NiceMan to a copy of this in the projects page.
Happy not dewing!
edit:
I meant to say. The saved GND plane filled ccts only show the outline of the GND plane polygon. To restore this GND plane infill just click on the Ratsnest (in the board design) in Eagle PCB. To Build either the 4 or 6 channel all up is from $91 to $113 Aust. This is very cheap for the specifications these ccts can supply.
Last edited by wasyoungonce; 11-08-2010 at 02:31 PM.
A 4 channel PCB design & 6 Channel PCB design. I have included a parts list and circuit description document. All documents are zipped for smaller size.
I'll be building a 6 channel PCB soon as I need the extra ccts.
I'll get NiceMan to a copy of this in the projects page.
Happy not dewing!
edit:
I meant to say. The saved GND plane filled ccts only show the outline of the GND plane polygon. To restore this GND plane infill just click on the Ratsnest (in the board design) in Eagle PCB. To Build either the 4 or 6 channel all up is from $91 to $113 Aust. This is very cheap for the specifications these ccts can supply.
None of the zipfile seemed to work for me. Tried them on two computers.
None of the zipfile seemed to work for me. Tried them on two computers.
Worked fine here.
Downloaded zip file, extracted to a folder and opened them with eagle PCB...no problems.
Patent office...yeah right patent on a 555cct...I laff .
I'll be revisiting these ccts soon. I'm plating with a stepper focuser controller and when done I'll update these probably with over current feed back control...hopefully.
But the principle of using a "high side switch" for dew heaters...is absolutely important. Circuit still going strong since built. I had to replace one potentiometer as I banged it against a wall and damaged it....my fault.
The design looks great but had a few questions with the basics.
With PWM, aren't you basically raising one terminal of the dew strap to +12 and then to zero with the FET? Not sure how the outer sheath can become +ve if the core is the terminal being raised high and outer sheath is connected to ground, won't that stay at ground?
FET's and transistors aren't my strong suit so pls pardon the ignorance.
Secondly, can't a single microcontroller be used if all that's needed is a TTL/CMOS high/low at the gate of the FET?
with an mcu, you can easily get 4 or more independent channels without actual pwm by just changing the interval between high and low pulses with the pot using ADC.
hikerbob's thread had a simple circuit with an 08m picaxe driving an FET. so can't that be extended to 4 or more channels?
what is the max voltage to be applied to the FET?
thanks
Just two comments
Re low side switching and shorting, if you use "unswitched" RCA connectors on the controller, you can always wire the 12V to the centre, and this makes it much more difficult to accidentally short out
I have attached a simple circuit i use based on an LM324 opamp
( as well as a finished unit to show its not complicated )
Three sections of the OPAmp make up a proper sawtooth output waveform, and the 4th section is used to provide a buffered power supply to drive the Mosfet.
By adding one more LM324, you can use the sawtooth waveform to feed 4 more supplies and hence get 5 independent channels out.
I would have to say tho, that after playing with multichannel devices like this, the cabling becomes a nightmare
What i am looking at doing is to make a very small PCB using SMT components, and actually build a controller into each dewstrap.
That way, all that is reqd is a simple 12V supply bus.
Andrew
Yes, SMT Picaxe on the strap is brilliant . But WHAT then are you doing posting that stupid archaic analog cct mess
, hence was always live, even when not connected.
and thanks for sharing, now to the pattern office 
.
. But WHAT then are you doing posting that stupid archaic analog cct mess 
