indulis

The best thing to use is de-ionized water if you can find some. Go to a good golf cart forum... they deal with this sort of stuff every day!! Check out http://www.golfcarcatalog.com/phpbb2/viewforum.php?f=2&sid=9d4161888160d5bbdf41fac46ef484ba

Yup......... your right!! I guess that's what happens when you get old and take a "stupid pill" in the morning!! I have not built this circuit, but I did try to simulate it in 3 different simulators, Workbench, Intusoft and MicroSim, with the same results every time. I couldn't get a output voltage greater than around 10.5V... may I ask what the origin of the circuit is??

That will never happen... read my post above in regards to the value of R1!!

Are you suggesting that the Norton & Thevenin Theorems somehow says that it makes a difference in this circuit???

More often than not, it's a matter of convienience when connecting the die to the carrier. You don't want the bonding wires running all over the place.

I am not a "RF" person, but to the best of my knowledge, FM (Frequency Modulation) is a "line-of-sight" transmission, in other words, if you drive behind a mountain, which blocks "line-of-sight" to the transmission antenna, you will loose the signal. If the signal got so weak that the receiver couldn't detect that it was stereo, I suppose you could look at that as receiving it as "mono" signal to a greater distance.

You don't... the FET's "share" the current so it doesn't have to be as large. If you don't like that thought, then solder the the "third" leg to a "bus bar".

Looks like D1 clamps the negative portion of the waveform that's AC coupled thru C4 and D2 is the rectifier. The only thing you have to keep in mind when selecting the diode is PVI (peak inverse voltage) and the current rating. Use a Schottky diode for a lower forward drop (better efficiency) and much less reverse recovery time at the expense of higher junction capacitance... if you don't care about stuff like that, use any old fast recovery with the proper ratings. What is the expected output current? Can't be much because of R1, the 1K (at 1mA of current that would be a 30V drop). You might consider a boost converter based on a 555 used as a PWM.

It's a series circuit... it makes no difference in which order you place the components. The same current flows through the ALL the circuit elements (battery, restistor, LED).

John, As I said before, from an analytical standpoint I agree you. No one disputes that under certain conditions, the 2N2219 will smoke, but that will ONLY happen when the pass transistors Hfe is at it's minimum. So, unless you have statistical data, that you can post, from the 2N3055 manufacturers that shows beta distribution is something other than "standard", you haven't proven a thing... it's just theory again. I couldn't find a 2N3055 data sheet that graphs Hfe at high Vce voltages (I looked at 6 different manufacturers)… know where I can find one??

Try measuring the current gain of a few hundred 2N3055 transistors to find one with a gain that's at it's specifications min level!!! I'm sure "gain variation" follows a standard distribution and components with spec's out beyond 6 sigma are hard to find!! I agree with you from a analytical perspective, but "theory" is just theory, until proven otherwise. Until such time that as many non-fuctioning supplies are built as functioning ones because of minimum beta of the pass transistor, "your theory" isn't supported by hard data, irreguardless of what "good design practices" tells us. You know... there a lot of places out there that design to "nomimal" specifications, because "statistical voodoo" says you can get away with it cost wise!!

From a purely analytical standpoint, "the theory expert" does make SOME valid points!! From a practical, "I only have to build one of these, not 10,000...", it probably doesn't matter!!

equivalent series resistance and equivalent series inductance

Here's probably more than you want to know about it... it shows a split primary winding of 13-13. This is because the program can't do half turns. TRANSFORMER PERFORMANCE SUMMARY Design Name: Designer: Date/Time: 06/30/06 09:45:45 Project: Untitled Notes: TRANSFORMER DESCRIPTION Core Family: ER Ferrite Core Weight (Grams): 1.800 Geometry: ER14.5/3/7 Total Gap (in): 311.7u Material Name: 3F3/200K400K_100C Spacer Thickness (in): 155.8u Manufacturer: PHILIPS Window Fill (%): 62.67 CORE DESCRIPTION Eff. Core Area (m^2): 17.60u Min. Core Area (m^2): 17.60u Winding Length (in): 0.1024 Winding Height (in): 0.1398 Avail. Window (m^2): 9.230u Area Product (m^4): 162.4p Min. Core Gap (in): 103.5u Volume (in^3): 20.38m Inside Diameter (m): 4.700m Surface Area (m^2): 580.0u Mean Length Turn (in): 1.020 Winding Shape: Round Max. Permeability: 2.400k Max. B, linear u (Gauss): 2.400k Sat. Flux Density (Gauss): 3.100k Res. Flux Density (Gauss): 1.100k Mean Mag. Path Len. (m): 19.00m TRANSFORMER PERFORMANCE DATA Flux Swing Type: half wave Input Waveform: pulse Duty Ratio (Pct.): 0.5000% Current Rise/Fall (Pct.): 0% Output Power (Watts): 0.6885 Magnetizing Ind. (Henry): 237.2u Pk. Flux Density (Gauss): 2.509k Core Loss (Watts): 0.2346 AC Flux Density (Gauss): 977.9 Copper Loss (Watts): 259.6u Ambient Temp. (deg C): 55.00 Core AwAc (m^4): 162.4p Temp. Rise (deg C): 28.75 Frequency (Hertz): 400.0k Volts/Turn: 2.754 USER DEFINED PERFORMANCE DATA Trise 28.75 Winding fill % 62.67 Tlevel 0 Total Weight (Pounds): 4.670m Copper Loss 259.6u Core Loss 0.2346 Bac(max) 100.0k Bac 977.9 Output Power 0.6885 Calculator 4.000 Gap 791.7u Lmag 237.2u Max Strands 2.000 Min. Turns 1.000 drive winding 1.000 Core weight 1.800 Efficiency 65.89 Round Coef. 0.9900 Winding pitch 1.000 Primary Turns 14.90 Idens(max) 100.0k Bp(max) 100.0k Vdrop2 4.725m Ploss 47.25u Vterminal 35.80 Jn 20.13 CuWt 0.2066 WINDING DESCRIPTION, RATINGS, AND CHARACTERISTICS Winding Number: 1 2 3 Primary or Secondary: pri sec pri Volts Specified (Volts): 35.80 12.43 35.80 Volts Average (Volts): 35.80 13.77 35.80 AC Current (Amps): 10.00m 50.00m 10.00m DC Current (Amps): 0 0 0 AC Resistance (Ohms): 0.2925 73.25m 0.4725 DC Resistance (Ohms): 0.1060 62.59m 0.1492 Power Loss, Copper (Watts): 29.25u 183.1u 47.25u Current Density (Amp/in^2): 129.9 828.0 129.9 Wire Type: HF HF HF Wire Size (AWG): 39 40 39 Wire Height (in): 4.331m 3.898m 4.331m Wire Width (in): 4.331m 3.898m 4.331m Wire Strands: 8 8 8 Turns: 13 5 13 Number of Layers: 6.500 2.500 6.500 Turns per Layer: 2.000 2.000 2.000 Start ID: 1 3 1000 Finish ID: 1000 4 2 Pitch: 1 1 1 Layer Insulation (in): 0 0 0 Wrapper Insulation (in): 3.000m 3.000m 3.000m End Margins (in): 0 0 0 Leakage Ind. Next (Henry): 1.296u 223.9n 0 Leakage(L->Sector) (Henry): 0 0 0 Leakage(Sector<-L) (Henry): 0 0 0 Winding Capacitance (Farad): 5.628p 9.147p 11.08p Capacitance to Next (Farad): 9.078p 9.979p 0 IR Drop 2.925m 3.663m 4.725m Copper Loss 29.25u 183.1u 47.25u Loaded Voltage 35.80 13.76 35.80 Current Density2 20.13 128.3 20.13 Winding Weight 98.78m 36.43m 0.1428 DESIGN CONSTRAINTS Max Window Fill (%): 100.0 Max Temp. Rise (deg C): 50.00 Max Pk. Flux Dens. (Gauss): 100.0k Max AC Flux Dens. (Gauss): 100.0k Max Cur. Dens.(Amp/in^2): 645.2k Waveform: square Auto Margin for Lead Exit: no Pitch (dia/turn): 1 K Conduction: 3.990 K Convection: 710.0 K Insulation: 2.000m K Dielectric: 3.000 Thermal Model Level: 0 Rac Method B&L SPICE MODEL AND SYMBOL NAMES Spice Model Name: not saved Spice Symbol Name: not saved

The transformer design isn't that bad... 25 turns primary and 5 turns secondary. For a SMPS, I think you could get away with a Feroxcube ER14.5 core with 3F3 material (inductance factor of 1400nH +/-25%) running at 400KHz on a single ended forward converter. It would have an ~ duty cycle of 46% (so reset wouldn't be a problem). Also the flux swing is limited to 2000 gauss (3F3 saturates around 3500 gauss @ 100C). Oh... this is all base on a 110VAC half wave rectified RMS voltage of~78V feeding the DC-DC.

If you look at the LM317HV spec, you will see that the max input to output voltage is 60V. That does not mean 60V is it!!! You could have 120VDC in and 100V out which is fine because you haven't violated the 60V max differential.

One guy here where I work does some "moonlighting" and has used, with some success, a small toaster oven to reflow solder paste on SMT PCB's. The only trick is to find the right temperature profile.

The "multiwatt15" package isn't really meant to go in a socket.

Use caution when talking about positive and negative voltages in regards to transistor biasing... use "potentials" so that it applies in all cases. see: http://www.electronics-lab.com/forum/index.php?topic=7097.msg43806;topicseen#msg43806

Avalanche breakdown occurs in all diodes!!! It's just a question of how well the breakdown voltage level is controlled (through doping levels). Reverse bias any "plain Jane" diode and it will work just like a zener, reverse the polarity across a zener, and it will behave just like a diode. The "zener" is just a "special case" of a "regular" diode.

Yes I did, and yes it is!!!! Those that know, know and those that don't, don't!!! The package is PADS (more than adaquate)... I don't do the layout's, the CAD designers do it with my guidance... I'm a circuit designer, not a PCB designer. Just in case.... I'm NOT talking about the auto-routers ability to connect all points in a node!!! I am talking about how it routes those connections. The exact path that the etch follows. The router that can take into account induced pickup due to a high field strength flux fields doesn't exist. In fact a "good" analog simulator for power analog dosn't exist, so forget about layout tools!!! Consider the subject closed!!!

This is not "my diagram".... The point I was/am trying to make is that shielding doesn't necessarily have to tie to "ground" or a "common point" of a circuit all the time, and sometimes, the shield is more effective if tied elsewhere, this is a irrefutable fact. The notion that all shields have to be tied to a "zero" or a "ground" or a "common reference" or whatever you wan to call it, is indeed a myth!!! Driven guards are the perfect example of this. The number one choice of where to connect a driven guard to is too a buffered version of the signal you are trying to guard. The number 2 & 3 places to connect the guard are the power rails or ground.... sorry, the supply rail's return. EMI energy can be diverted to anywhere (preferabaly to it's point of origin) as long as it stops interfering with a circuits operation. Whether or not is easy to do, as far as having to keep track of which shield is connected to what, is irrelevant..

I don't want to insult anybody

The "common reference" can be anything I want it to be, it DOES NOT have to be something called ground or zero... just a "common point". In this circuit for example, you could take the +5 from back wheel battery, connect the +9 and the 5V return from the blade motor and call this my reference. It is a myth that shields have to be connected to ground or a single common to be effective in "protecting" signal integrity... If what your saying is 100% correct, then this guy.... http://www.ieee-houston.org/Ads/DonWhite/DonWhite.htm ... is 100% wrong. I attended a couple of his seminars years back down in Florida (it was February... lucky me) for "Grounding and Shielding" and "EMI control in backplanes". Actually... now that I think about it, it was quite a while ago...1984 Anyway... these are 3 floating circuits... why would I "have to" tie the "commons" together?? I can define the supply for the DC motor as being -5V, so I would be tying the -5 return to the common. It might be found that it works "best" when the microcontroler signals are shielded with +9V..... when it comes to EMI and EMC, you OFTEN have to "think-outside-the-box"!!!

Yes, but your talking digital. Digital is a different story... yes, it'll work there. I was talking power analog design. Do you mean this software!!! http://www.cheap-software-megastore.com/index.php?target=desc&progid=1329