Post by thread:[PIC]: RC oscillator

Hi, I have an application which requires a an RC oscillator. Since the design will be subjected to high vibration in the field, it is felt that a crystal or resonator would be too fragile. Right now we are operating at 4MHz, but this is may end up at 1 MHz or below. The problem is that when we test the oscillator over temperature it moves several percent between 25C and 50 C, at which point it can no longer maintain serial communications because the clock has drifted too far. This is unexpected because the we are using top quality high stability resistors and capacitors, which should only move about 20-30 PPM/C each. I thought I had seen graphs of the stability of the RC mode from device to device and over temperature, but I can't locate these in the datasheet or app notes. I would like to know if anyone out there has experience with the RC oscillators performance over extended temperature(we'd like to go up to 125C). Hopefully I could get a few suggestions to stabilize our design. Thanks in advance for any help Don McManus Bently Nevada Corporation

At 08:42 AM 5/18/00 -0700, you wrote: > I thought I had seen graphs of the stability of the RC mode from >device to device and over temperature, but I can't locate these in the >datasheet or app notes. I don't think that the RC mode is suitable for serial communications, not in a production product that has to operate over a wide temperature range. Why don't you put an SMT or a 20 cent leaded resonator on there? You will get guaranteed acceptable performance (without a lot of margin, but O.K.) Resonators are much more robust than crystals. Only a percent or two change and you will have decreased serial noise immunity, at 5% or so it typically will stop working. I have used LC oscillators on PIC's to improve the guaranteed accuracy, but the accuracy certainly not good enough for this application. The stability, OTOH, might be just ok for asynchronous serial comms. Best regards, =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= Spehro Pefhany --"it's the network..." "The Journey is the reward" spam_OUTspeffTakeThisOuTinterlog.com Info for manufacturers: http://www.trexon.com Embedded software/hardware/analog Info for designers: http://www.speff.com Contributions invited->The AVR-gcc FAQ is at: http://www.bluecollarlinux.com =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=

> I have an application which requires an RC oscillator. Since the >design will be subjected to high vibration in the field, it is felt that a >crystal or resonator would be too fragile. Don, RC oscillators can be made stable enough for controlling serial async data transfer but in your circumstances I would first consider whether a resonator or crystal is really as fragile as you think it is. How strong is a capacitor compared to a resonator? What sort of environment are you in? Properly mounted a resonator or crystal will survive massive accelerations. I don't know details on what they use but the military fire GPS units in artillery shells (!) Unless you are using this equipment to tamp in your sticks of gelignite you are unlikely to get higher accelerations than that. Equipment with survivable accelerations of in the order of 1000 g is quite standard. Close attention to mounting detail and massive common sense seems to cover most of it. I don't know about the internal mounting structures in crystals but resonators "seem" to be rather more monolithic. You could try an easy test. Mount a resonator on a steel block - hold in place with epoxy and press firmly into place against the steel as the epoxy sets. Drop steel block from various heights onto paper sheet on steel plate on concrete floor (resonator uppermost :-)) from various heights. If one assumes about zero compression of the steel-steel interface any deceleration is carries out over the thickness of the paper (or less as it will have some compression thickness). Acceleration in g's = Fall distance/Stopping-distance eg for a 0.010" = 0.25mm = 0.00025m thick stopper A drop from 250mm results in a deceleration of 0.250/0.00025 = 1000g IF this kills your resonator try embedding it in a very small amount of stiffish padding and try again. The g forces will be MUCH reduced (as seen above) by allowing it even a small amount of "compliance". regards Russell McMahon _____________________________ >From other worlds - http://www.easttimor.com http://www.sudan.com What can one man* do? Help the hungry at no cost to yourself! at http://www.thehungersite.com/ (* - or woman, child or internet enabled intelligent entity :-))

Chris A wrote... >Is there a rule-of-thumb to be used when using RC on F876's? The >book says 3K - 300K etc, but how do I know the frequency? Can >I just use T=RC? The PIC Mid-Range Reference Manual (and many of the device data sheets too, if I recall correctly) gives some data on RC oscillator performance. Table 31-1 on page 628 of the MRRM (document DS31031A) gives oscillator frequency as a function of various resistor/capacitor values. Dave -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email listservKILLspammitvma.mit.edu with SET PICList DIGEST in the body

>>Is there a rule-of-thumb to be used when using RC on F876's? The >>book says 3K - 300K etc, but how do I know the frequency? Can >>I just use T=RC? >The PIC Mid-Range Reference Manual (and many of the device data sheets >too, if I recall correctly) gives some data on RC oscillator >performance. Table 31-1 on page 628 of the MRRM (document DS31031A) >gives oscillator frequency as a function of various resistor/capacitor >values. Jinx used to have curves on his website IIRC that were frequencies he measured with a 16F84 series device. This should not be too different to any of the other 16... series devices at similar voltage to what he used. I cannot remember the URL unfortunately. -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email .....listservKILLspam.....mitvma.mit.edu with SET PICList DIGEST in the body

I just realised that if you were asking about RC speeds then you probably don't have a way of measuring the frequency directly. The simplest way I can think of to do it indirectly would be to set up a timer IRQ and measure how long it takes for a certain number of IRQs. eg if the RC you choose gets you 1MHz, then that will be approx 1000 IRQ/sec (= 1,000,000/4/256). Using a pre-scaler and counters to extend the measurement period and a stopwatch should give you a pretty good idea of what the actual RC frequency is. It'll be as close as you could expect with T & V dependent RC anyway -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email @spam@listservKILLspammitvma.mit.edu with SET PICList DIGEST in the body

> I thought that there would be a rule-of-thumb to calculate > these things Not that I'm aware of, but maybe you could deduce a formula from my chart. It's reasonably straight-line. Not that I bothered, sorry. Didn't need to. There is some variance between different PICs (both individually and families), as you'll see from the loose specs in the data sheet (ISTR 7% ?) and tolerances of caps and resistors plus temperature fluctuations. If you need a 100ms delay, work out how many instruction cycles to do that, that is both in the internal 00-FF loop and the external counter incrementation So a basic loop takes 1 + 256 +1 + (256 * 2) = 770 cycles (did I get that right ? Been a very very busy day) clrf count1 (1) loop incfsz count1,f (1) or (2) goto loop (2) At 4MHz this would take 770us. Bump another counter 128 times (just test its bit7) to give you the 100ms, give or take the "incf" and "goto loop" and test for the second counter byte, which will be a small fraction of the overall time But you still need, I think, to set up a timer IRQ routine as I suggested to actually measure the frequency if you have no scope or freq counter -- http://www.piclist.com hint: To leave the PICList RemoveMEpiclist-unsubscribe-requestTakeThisOuTmitvma.mit.edu