Post by thread:Write voltage and MAX232

> The idea consist to use the charge pump of the +12 Volt of the MAX 232 to > sink voltage and current to use for programming the 16F84 when only five > volt are present. I hope you mean source, and not sink.. Charge pumps don't sink current. That said, I have no idea how well regulated that output is, or how dependent it is on what's connected to the serial outputs. I've used the + and - outputs to power op-amps, and actually it works quite well, but opamps dont' care much about supply voltage. 1V of ripple, and a couple of volts of uncertanity wouldn't bother my designs at all. I suspect the VPP spec for the 84 is a bit tighter than that. You might find that it goes high enough to use a 78L12 to regulate it, I wouldn't suggest a zener, that would cause the charge pump to work overtime. (The zener draws more current with higher input, the linear draws same current.)

Dave VanHorn wrote: > > > The idea consist to use the charge pump of the +12 Volt of the MAX 232 to > > sink voltage and current to use for programming the 16F84 when only five > > volt are present. > > <snipped> > > suspect the VPP spec for the 84 is a bit tighter than that. > Vpp has to be 12-14V with 0.25V resolution. The HV used is generated on-chip, so basically NO current is used during programming, from Vpp. (50 mA from Vdd, though, per the specs)... I've thought of using a charge pump based TLC555 design in the past, I think a MAX232 would work, provided you made sure you had a stable 12-14VDC there. Watch the rise time & so forth requirements, they're pretty well-enforced <G> Mark, spam_OUTmwillisTakeThisOuTnwlink.com

cacciavite@bigfoot.com wrote: > The idea consist to use the charge pump of the +12 Volt of the MAX 232 > to sink voltage and current to use for programming the 16F84 when only > five volt are present. You seem to be under a misapprehension. The MAX232 doesn't *have* a +12V rail. It contains a voltage doubler and a voltage inverter; it doubles 5V to 10V, then converts +10V to -10V. You could only get 20V out of it by floating the effective "ground" rail of the PIC from the -10V, which means you suddenly need a number of level translators. Perhaps an LM324? And regulated, it ain't! OK? -- Cheers, Paul B.

> Van: .....cacciaviteKILLspam@spam@bigfoot.com <leo.depaloKILLspamPOMETIA.IT> > Aan: .....PICLISTKILLspam.....MITVMA.MIT.EDU > Onderwerp: Write voltage and MAX232 > Datum: zondag 8 november 1998 21:46 > > Hi PICer, > > I have been an idea about the voltage level to program the PIC 17F84 and I > need to discuss with this list you about the faceability. Ahem .... I believe you mean the _16_ F84 ? > The idea consist to use the charge pump of the +12 Volt of the MAX 232 to > sink voltage and current to use for programming the 16F84 when only five > volt are present. My spec's of the MAX232 say that with a load of 3KOhm ( ! ) the voltage drops to about 8 volts. The maximum voltage (NO load present) is 10 volts. I'ts somewhat short of the 12.5 volts you need for EPROMS. For a 16?84 (or other EPROM / FLASH controllers) it's O.K. , because the voltage is only used as a "go into programming-state" signal. It will draw _very_ little current. My own experiences showed that 9 volts was enough to put the 16C84 into the "programming-state". Greetz, Rudy Wieser

The other side of the moon... It is easy to apply ADP111x/ADP3000 chip to obtain +12v from +5v with only few of additional C, L, 1N4148 & R's. ADP111x/3000 == LT111x/3000 (AnalogDev's or LinearTech) WBR Dmitry. > > Hi PICer, > > I have been an idea about the voltage level to program the PIC 17F84 and > I need to discuss with this list you about the faceability. > > Ahem .... I believe you mean the _16_ F84 ? > > The idea consist to use the charge pump of the +12 Volt of the MAX 232 to > > sink voltage and current to use for programming the 16F84 when only five > > volt are present.

On Mon, 9 Nov 1998, Paul B. Webster VK2BZC wrote: > You seem to be under a misapprehension. The MAX232 doesn't *have* a > +12V rail. It contains a voltage doubler and a voltage inverter; it > doubles 5V to 10V, then converts +10V to -10V. That is true, HOWEVER it has got square wave on some of the external pump caps, and nowhere does it say, that you are not to connect another small cap to it, and 2 shottky diodes or a 4066 in active rectifier disguise to boost another ~5V externally. This needs to be elaborated but it will work. The question is, why are you trying to do such impossible things when a 3:1 ferrite transformer driven by the open drain PIC pin can do all you want and more ? Ok, you can use a VFET to boost the power, but it may not be necessary. Peter

The question is, why are you trying to do such impossible things > when a 3:1 ferrite transformer driven by the open drain PIC pin can do all > you want and more ? Ok, you can use a VFET to boost the power, but it may > not be necessary. > > Peter This would indeed be a neat trick. He's trying to generate programming voltage for writing the PIC. I suspect that this may cause instability in the voltage source. :) Handling the flyback pulse could be interesting as well. Can we spell CMOS latchup? I think that FET would be well advised. I've heard processors discussed as SMPS controllers before, but never seen it implemented.. That's interesting, because they have the potential to be a lot more intelligent than existing analog chips. External drivers are not a problem, but driving the drivers might be. The CS3844 for example, provides 1A in either direction to rapidly charge/discharge the switching FET's gate capacitance. Any processor is going to have a hard time with that!

On Mon, 9 Nov 1998, Dave VanHorn wrote: > The question is, why are you trying to do such impossible things > > when a 3:1 ferrite transformer driven by the open drain PIC pin can do all > > you want and more ? Ok, you can use a VFET to boost the power, but it may > > not be necessary. > > > > Peter > > This would indeed be a neat trick. He's trying to generate programming > voltage for writing the PIC. I suspect that this may cause instability > in the voltage source. :) I was thinking of using a PIC to program a PIC of course. This is a cheat on bootstrapping ;) > Handling the flyback pulse could be interesting as well. Can we spell > CMOS latchup? > I think that FET would be well advised. The OD PIC pin can switch a coil at 5V Vdd without the latchup becoming dangerous (of course you use a zener or diode to limit it just in case). > I've heard processors discussed as SMPS controllers before, but never > seen it implemented.. That's interesting, because they have the ;) About 50% of the camcorders on the market use the PWM outputs of a processor to control a pair of choppers directly. These drive the drum and capstan motor, with the control loops closed through the resp. sensors and the processor algorythm. There are several schematics of special purpose DC/DC converters built with PIC16C71/711s, with four channels each and serial control. Need I go on ? > potential to be a lot more intelligent than existing analog chips. > External drivers are not a problem, but driving the drivers might be. > The CS3844 for example, provides 1A in either direction to rapidly > charge/discharge the switching FET's gate capacitance. Any processor is > going to have a hard time with that! As long as we talk small, a PIC output can drive a small VFET (like a BSS175 etc) directly properly and this leads to outputs of the order of 0.8 A per channel. Driving VFETs too fast without a reason is painful in the RFI domain. You almost always end up trying to slow it down afterwards. Peter

> > Handling the flyback pulse could be interesting as well. Can we spell > > CMOS latchup? > > I think that FET would be well advised. > > The OD PIC pin can switch a coil at 5V Vdd without the latchup becoming > dangerous (of course you use a zener or diode to limit it just in case). Is this specified, or just something that seems to work? > There are several schematics of special purpose DC/DC converters built > with PIC16C71/711s, with four channels each and serial control. Need I go > on ? I didn't say it couldn't be done, just that I've not seen it done. SMPS I mean, not servo control for a motor. There are strong similarities though. > As long as we talk small, a PIC output can drive a small VFET (like a > BSS175 etc) directly properly and this leads to outputs of the order of > 0.8 A per channel. Driving VFETs too fast without a reason is painful in > the RFI domain. You almost always end up trying to slow it down > afterwards. Easier to slow down than to speed up. Excessive dV/dT is always bad though.

On Mon, 9 Nov 1998, Dave VanHorn wrote: > > > Handling the flyback pulse could be interesting as well. Can we spell > > > CMOS latchup? > > > I think that FET would be well advised. > > > > The OD PIC pin can switch a coil at 5V Vdd without the latchup becoming > > dangerous (of course you use a zener or diode to limit it just in case). > > Is this specified, or just something that seems to work? The RA4 pin can be raised to +12V without damage as far as I know. So, if you connect a coil from +Vdd to RA4 and a Z10V8 from RA4 to GND you should be reasonably safe. I am quite certain that the manufacturer never intended this to be used like this but I did not ask him ;). Peter

> > Is this specified, or just something that seems to work? > > The RA4 pin can be raised to +12V without damage as far as I know. So, if > you connect a coil from +Vdd to RA4 and a Z10V8 from RA4 to GND you should > be reasonably safe. I am quite certain that the manufacturer never > intended this to be used like this but I did not ask him ;). > > Peter Danger! Will Robinson.. I would want to look at that very carefully, the energy stored in that coil has to dump somewhere, and it dosen't sound like this is a good place. A 5.1 zener would be a safer bet, as long as VCC is well regulated. (1%) (I just can't shake my mentality of "it's got to be something that will work for 100k units, not just one)

Peter L. Peres wrote: >> You seem to be under a misapprehension. The MAX232 doesn't *have* >> a +12V rail. It contains a voltage doubler and a voltage inverter; > That is true, HOWEVER it has got square wave on some of the external > pump caps, ... The question is, why are you trying to do such > impossible things ... There's a lot to be said for the old TL497 or various other similar devices. If however what you have on hand is a MAX232 and an LM324, I think "Yet Yet Another PIC16F84 Programmer" could be concocted. And I sure as eggs am *not* going to try it. -- Cheers, Paul B.

On Mon, 9 Nov 1998, Dave VanHorn wrote: > Danger! Will Robinson.. > I would want to look at that very carefully, the energy stored in that > coil has to dump somewhere, and it dosen't sound like this is a good > place. The RA4 pin is Open Drain. It is also specced somewhere (I need to check) to be raise-able to within 0.6V of Vpp or such. The Zener in parallel with the pin to GND will limit any excursion up or down at the pin. The other end of the coil was connected to a 4.8V NiCd pack with an internal R of less than 0.2 ohms, bypassed by a large cap and by a reversing protection 5V6, plus an 800 mA slow blow fuse. I don't normally do stupid things. Peter

> > > Is this specified, or just something that seems to work? > > > > The RA4 pin can be raised to +12V without damage as far as I know. So, > if > > you connect a coil from +Vdd to RA4 and a Z10V8 from RA4 to GND you > should > > be reasonably safe. I am quite certain that the manufacturer never > > intended this to be used like this but I did not ask him ;). > > > > Peter > > Danger! Will Robinson.. > I would want to look at that very carefully, the energy stored in that > coil has to dump somewhere, and it dosen't sound like this is a good > place. > A 5.1 zener would be a safer bet, as long as VCC is well regulated. (1%) > (I just can't shake my mentality of "it's got to be something that will > work for 100k units, not just one) > But then the max voltage obtained would be a little over 5 volts which defeats the whole purpose of the exercise which is to generate enough voltage to program a pic/eprom. The trouble with zeners is that the common or garden variety aren't very good in this kind of application. They are slow and have quite a high dynamic impedance. Large, fast current spikes can cause the voltage to rise appeciably. If the output from the switcher is suitably loaded then the zener should never have to work very hard, it's only wasting power if it is. Regards Mike Rigby-Jones EraseMEmrjonesspam_OUTTakeThisOuTnortel.co.uk

Hello Peter. What the way you put into limiting of maximal current through RA4 pin ? You may easy damage or overheat the PIC simply by overloading the RA4 pin with current bigger than maximal current is allowed(according specs). This is very important 'cause the coil save the current not the voltage. BTW... This mulfunction can be found unexpectively many time later. WBR Dmitry. > The RA4 pin is Open Drain. It is also specced somewhere (I need to check) > to be raise-able to within 0.6V of Vpp or such. The Zener in parallel with > the pin to GND will limit any excursion up or down at the pin. The other > end of the coil was connected to a 4.8V NiCd pack with an internal R of > less than 0.2 ohms, bypassed by a large cap and by a reversing protection > 5V6, plus an 800 mA slow blow fuse. I don't normally do stupid things. > > Peter

Am 9 Nov 98 um 14:12 hat Dave VanHorn geschrieben: > > > Handling the flyback pulse could be interesting as well. Can we spell > > > CMOS latchup? > > > I think that FET would be well advised. > > > > The OD PIC pin can switch a coil at 5V Vdd without the latchup becoming > > dangerous (of course you use a zener or diode to limit it just in case). > > Is this specified, or just something that seems to work? Deeply buried and hidden in Microchips documenation, there's a "up to 12V" spec. It's in the thick "midrange manual", if memory serves me right, I haven't found it anywhere else. -- o ( Wolfgang.Stroblspam_OUTgmd.de (+49 2241) 14-2394 /\ * GMD mbH #include _`\ `_<=== Schloss Birlinghoven, <std.disclaimer> __(_)/_(_)___.-._ 53754 Sankt Augustin, Germany ________________

On Tue, 10 Nov 1998, Dmitry Kiryashov wrote: > Hello Peter. > > What the way you put into limiting of maximal current through RA4 pin ? The coil and the Ton on RA4 and the watchdog in the PIC ;) > You may easy damage or overheat the PIC simply by overloading the RA4 > pin with current bigger than maximal current is allowed(according > specs). This is very important 'cause the coil save the current not the > voltage. If you had tried to test PICs to destruction like this to see what gives, you would know that the main failure mode is by overheating in this case, which may take a few days or more to do any damage... Of course I do not endorse this but I did use a PTC instead of a fuse once, and put it in contact with the chip body for such purposes. Peter

|I was thinking of using a PIC to program a PIC of course. This is a cheat |on bootstrapping ;) Having designed and built a few PIC programmers and had mostly good luck but a few problems as well, I would highly recommend that hobbyists plan on using a "dirtball" programmer to produce a 16x84 which can then be used in a more advanced programmer. Even though the 16x84 programming specs are very much like those of the other serially-programmed PICs and it would seem that a device which can do one should be able to do them all, the reality is a bit more complicated: the non-84 PICs are much more picky when it comes to their programming specs. AFAIK, there is no way to do any real damage to a 16x84 when you're trying to program it unless you drive one of the pins outside the "absolute maximum" rating; by contrast, other PICs may be damaged by overly-long program pulses and other such annoyances. For this reason, while I have no qualm with 16x84 programmers that have the PC generate all the clock and data timings, I would not recommend that this approach be used with the other types of PICs. When running straight DOS, with interrupts disabled, a PIC can do a decent job of producing the necessary timings, but under Windows, Linux, etc. it's possible for the CPU to lose its timeslice between the "Start programming" and "End programming" commands. If this happens very often or for very long, it's possible to blow a column driver on the chip and be permanently unable to program any more bits on that column (on the 12C508, if memory serves, this would translate into one bit per every 16th word--if you were really desperate you could probably code around it, but why...)

Am 11 Nov 98, um 2:30 hat Mark Willis geschrieben: > RA4 pin max voltage is 14V folks, not 12V. It's in the .PDF files, do > a simple text search for Vpp or Programming & you'll find it. I haven't > seen any chip with an open drain RA4 that isn't 14V specced. Uh, oh, you're right, of course. It's on page 573 in the midrange manual, section 30, Electrical specifications. Voltage on RA4 with respect to Vss ............................... 0 to +14V" > > 12V is pretty well guaranteed to be safe, of course <<G> Well, in the 16F8X manual, there is NO such guarantee. On page 72 of 30430b.pdf, it says Voltage on MCLR with respect to VSS (2) .......................... -0.3 to +14V Voltage on any pin with respect to VSS (except VDD and MCLR) ..... -0.6V to (VDD + 0.6V) I don't know whether this is an oversight, or intentional. Does anybody know? <nofill> -- o ( KILLspamWolfgang.StroblKILLspamgmd.de (+49 2241) 14-2394 /\ * GMD mbH #include _`\ `_<=== Schloss Birlinghoven, <std.disclaimer> __(_)/_(_)___.-._ 53754 Sankt Augustin, Germany ________________

Am 11 Nov 98, um 2:30 hat Mark Willis geschrieben: > RA4 pin max voltage is 14V folks, not 12V. It's in the .PDF files, do > a simple text search for Vpp or Programming & you'll find it. I haven't > seen any chip with an open drain RA4 that isn't 14V specced. Uh, oh, you're right, of course. It's on page 573 in the midrange manual, section 30, Electrical specifications. Voltage on RA4 with respect to Vss ............................... 0 to +14V" > > 12V is pretty well guaranteed to be safe, of course <G> Well, in the 16F8X manual, there is NO such guarantee. On page 72 of 30430b.pdf, it says Voltage on MCLR with respect to VSS (2) .......................... -0.3 to +14V Voltage on any pin with respect to VSS (except VDD and MCLR) ..... -0.6V to (VDD + 0.6V) I don't know whether this is an oversight, or intentional. Does anybody know?

On Wed, 11 Nov 1998, John Payson wrote: > a decent job of producing the necessary timings, but under Windows, > Linux, etc. it's possible for the CPU to lose its timeslice between In the case of a (temporarily) dedicated linux box, the only reason you can 'lose timing' is because you are a boat anchor at kernel mode programming. Been there, and got over it. After all, you have the source and you can do with the scheduler whatever you want. Also temporarily locks disk access and other useful things. Peter