Strange MOSFET driver behaviour - Microchip TC4426
in [Design]
|
Prev: Please, DO NOT forget the only thing that matters on this planet........................……..
Next: Adios, amigos...
From: Jan Panteltje on 11 May 2010 10:21 On a sunny day (Tue, 11 May 2010 07:57:00 GMT) it happened nico(a)puntnl.niks (Nico Coesel) wrote in <4be90d8d.1464277890(a)news.planet.nl>: >>Yep, stay clear of Microfip. > >This sounds like the Pope converted to the Islam. Everything allright? Well, 'three strikes = out' policy. The 16F648 had EEPROM programming problems. The 16F690 sets in some mode the PWM output to 1 100% of the time for PWM angle zero (any other angle works OK), MOSFET killer. The 18F14K22 has that output problem at 64 MHz, but that is fixable, but also SPI does not work *at all* and I2C sucks too it seems (have not tried that one yet). In my view Microchip does not test their stuff for their own specs. Sort of leaves that to the user. I think that is really bad practice. Just because I did a few Microchip PIC projects does not mean I work for them :-) Microchip PICs have their use, I used those, and still use those in many projects. But I am not going for 'yet an other experiment' with a different model. Next project, if there is one, I may try some other thing. Maybe even do my own micro in FPGA, there are? a lot of open source cores on opencores.org, and then, if I need SPI, shift register is easy in FPGA :-) especially in the Xilinx ones. I still think that Microchip PICs are great and you can have more then one on one board, but you have to *pic* the right one for the job, and that seems to be trial and error. All has it's pros and cons, knowing what tool to use for what is the art is not that your own words? 'If I had a hammer' Experience. I did read Microsoft is stopping with Microsoft groups (on Usenet). There are a lot of those groups. probably everybody was complaining, so now will only they have their own (likely censored) forum. Usenet is a medium of truth, I can only write what I experience. What I would like is a small DIL FLASH based FPGA at low cost. The other thing is modules, those get cheaper, a small module with some Linux on it, a processor with a real stack, programmable with gcc, ethernet, USB, some I/O, may drop in price so low that it beats putting a board together yourself. Now that on a small .1 inch spaces header... But, I am working on an other 18F14K22 project, after all I have plenty of those laying around ... hehe Else a cheap PC mobo is only 50 Euro.... with onboard graphics. There you are. But that cannot beat, but only augment something like his: http://panteltje.com/panteltje/pic/freq_pic/index.html Everything has its use.
From: Spehro Pefhany on 11 May 2010 11:09 On Tue, 11 May 2010 14:21:58 GMT, Jan Panteltje <pNaonStpealmtje(a)yahoo.com> wrote: >But that cannot beat, but only augment something like his: > http://panteltje.com/panteltje/pic/freq_pic/index.html >Everything has its use. It would be interesting to see how the 18FxxKxx would do in that circuit. I'm guessing >150MHz typically.
From: Nico Coesel on 11 May 2010 11:41 Jan Panteltje <pNaonStpealmtje(a)yahoo.com> wrote: >On a sunny day (Tue, 11 May 2010 07:57:00 GMT) it happened nico(a)puntnl.niks >(Nico Coesel) wrote in <4be90d8d.1464277890(a)news.planet.nl>: > >>>Yep, stay clear of Microfip. >> >>This sounds like the Pope converted to the Islam. Everything allright? > >Well, 'three strikes = out' policy. >The 16F648 had EEPROM programming problems. >The 16F690 sets in some mode the PWM output to 1 100% of the time >for PWM angle zero (any other angle works OK), MOSFET killer. >The 18F14K22 has that output problem at 64 MHz, but that is fixable, >but also SPI does not work *at all* and I2C sucks too it seems (have not tried that one yet). > >In my view Microchip does not test their stuff for their own specs. >Sort of leaves that to the user. >I think that is really bad practice. > > >Just because I did a few Microchip PIC projects does not mean I work for them :-) > >Microchip PICs have their use, I used those, and still use those in many projects. >But I am not going for 'yet an other experiment' with a different model. > >Next project, if there is one, I may try some other thing. >Maybe even do my own micro in FPGA, there are? a lot of open source cores >on opencores.org, and then, if I need SPI, shift register is easy in FPGA :-) >especially in the Xilinx ones. I have been looking into that direction as well. The problem with most cores is that they are large and for some obscure reason try to be instruction cycle accurate. For what reason? It just takes extra space and costs performance. I have been tinkering with the idea to create an MSP430 compatible CPU core. That is a 16 bit core (small) with a very nice instruction set and there is a decent GCC compiler available for it. >What I would like is a small DIL FLASH based FPGA at low cost. > >The other thing is modules, those get cheaper, a small module with some Linux on it, >a processor with a real stack, programmable with gcc, ethernet, USB, some I/O, may drop in price >so low that it beats putting a board together yourself. >Now that on a small .1 inch spaces header... There are many companies that sell such modules for very low prices. -- Failure does not prove something is impossible, failure simply indicates you are not using the right tools... nico(a)nctdevpuntnl (punt=.) --------------------------------------------------------------
From: Tim Williams on 11 May 2010 14:04 "Didi" <dp(a)tgi-sci.com> wrote in message news:50303c5d-760e-4962-aec1-a22b31ae62cb(a)l31g2000yqm.googlegroups.com... > I am not sure I can do that on a ferrite core, at 500 kHz I manage > somewhat above 10 (I think I tested at 12) W through a 10mm outer > diameter, 3.8mm high toroid. It does get hot at this power though. I had made a ~3W flyback converter at 500kHz with a #52 core, and it got noticably hot (20-30C temp rise?). I don't recall what Bmax it used... http://webpages.charter.net/dawill/tmoranwms/Circuits_2010/Fast_DCDC.png This is the circuit. Silly me, green-blue is #52, not MPP. 15V applied to 15T over 1us is 1uWb/t. Hmmm... T50-52D seems to be the core, though A_L is different from my measurement. Ae = 22.3mm^2, so Bmax = 45mT. (Which is 476 At/m at mu_r = 75, and l_e = 31.9 mm, so MMF = 15At. Ipk was around 1A, so that's close enough.) On the other hand, I also made a 1MHz forward converter using a fairly small ferrite E core: http://myweb.msoe.edu/williamstm/Images/Fast_Forward3.jpg I was thinking 0.1T peak, but for size reasons, I had to go with more like 0.3T peak, which is typical of lower frequencies. You'd think hysteresis or lossy permeability would cause problems up there. Material is probably ordinary mu = 2000 stuff, not #43. Even so, it didn't heat up noticably. As far as I can tell, ferrite is practically lossless. Do whatever you want to it, as long as it's under 0.4T. Maybe doesn't apply so well at 10MHz+, but that's getting into RF anyway. > Can you make a guess how much there is to gain by using MPP > (presumably the lowest possible loss)? (right now my efficiency is > around 60% to 70%, I would guesstimate 65-70% as far as the inductor > is concerned). Hmm, continuing the earlier example, #52 is rated as 1.5W/cm^3 at 500kHz, 250 gauss peak (= 500G p-p). The core is 0.711 cm^3, so it should burn around 1.07W, which sounds about right. As for MPP, aha! They do indeed have an MPP material, -M125, in the 200C Series. Magnetics Inc. shows 60 mu MPP having about 70mW/cm^3 at 300kHz for the same flux density. At 500kHz, the formula evaluates to 0.18W/cm^3, so it's about 8 times lower loss. That's not too bad. I should go buy some. Tim -- Deep Friar: a very philosophical monk. Website: http://webpages.charter.net/dawill/tmoranwms
From: Didi on 11 May 2010 16:58
On May 11, 9:04 pm, "Tim Williams" <tmoran...(a)charter.net> wrote: > .... > As for MPP, aha! They do indeed have an MPP material, -M125, in the 200C > Series. > > Magnetics Inc. shows 60 mu MPP having about 70mW/cm^3 at 300kHz for the same > flux density. At 500kHz, the formula evaluates to 0.18W/cm^3, so it's about > 8 times lower loss. > > That's not too bad. I should go buy some. Oh, that's more than I would have hoped for. More than twice more, I guess :-). I'll have to give them a try as well. Thanks for the insight! Dimiter |