Replies to Minipulse Plus

[6666]

Thanks Qiaozhi

[sinclairuser]

Hi Qiaozhi

if you dont mind I have a hypothetical question for you about the preamp in the MPP, I'm working on another project
and the preamp would be handy, the question is , seeing that the preamp is AC coupled and does not have any offset
adjustments , do you strictly have to supply regulated + and - 5 volts to it, could you simply use
just two 9 volt batteries ?

thats how its done in a cscope tr200 i have that was my grandads, it makes for a simpler circuit but works well.

[6666]

thats how its done in a cscope tr200 i have that was my grandads, it makes for a simpler circuit but works well.

Thanks for the info, anybody got a schematic for that MD ?

[6666]

Hi Davor
After I run your 4093 sim a whole bunch of wave forms automatically display, what is the correct way to stop that from happening ? thanks.

[Qiaozhi]

Hi Davor
After I run your 4093 sim a whole bunch of wave forms automatically display, what is the correct way to stop that from happening ? thanks.

That's because the plot display has been saved. You can either delete the waveforms manually and re-save, or simply delete the ".raw" file.

[6666]

Deleted the raw file but waveforms are still generated.

http://s27.postimg.org/629ohve43/3093_5.jpg

[Qiaozhi]

Deleted the raw file but waveforms are still generated.

http://s27.postimg.org/629ohve43/3093_5.jpg

Sorry, that was incorrect. It's the ".plt" file you need to remove.

[Davor]

Thanks George, I assumed it will be easier to include a plt for easier identification of what's going on.

[6666]

Sorry, that was incorrect. It's the ".plt" file you need to remove.

Thanks that fixed it.

[6666]

Hi Qiaozhi

Do you find that if you are useing 12 volts and 300uH coil , that if you go much above the 58 - 60uS TX pulse width then the 740 starts to break down / avalanche ?

[Qiaozhi]

Hi Qiaozhi

Do you find that if you are useing 12 volts and 300uH coil , that if you go much above the 58 - 60uS TX pulse width then the 740 starts to break down / avalanche ?

The avalanche breakdown voltage for the IRF740 is 400V. Since the flyback is currently around 350V, then I wouldn't be surprised if that's true.

[scrungy_doolittle]

I can't convince a pic, or any other micro for that matter, to provide a continuously variable delay in a 10us range. I gave already a slightly different solution for this problem using a micro that varies a different delay which is not this tight. So far I'm not that much interested in micros.

Been a while since I was on, so exactly what kind of delay do you want? Given 16 mhz clock, a delay can be pretty simple if you just want to delay a signal. One pin in, one pin out. lets say you wanted a .6 usec delay. There are a number of ways todo this. One is to read the input pin, and or it with the msbit of a byte that you shift left first. Byte starts with all zeros, and implements an 8 bit shift register Basically it looks like this

depending on the clock speed and what the compiler generates, this is one way to do it. If it take 200 ns for the code to execuit, you would have a 1.6 used delay It can be
even less is you use a smaller number of bits. This is based on an 8 bit byte
do
{
if input pin is 1
delay_line=delay_line >> 1 | 0x80;
else
delay_line=delay_line >>1 & 0x7F;
ouput=delay_line & 1;
}while(1);

If you wanted 4 bits of delay, your output would be changed to this
output=delay_line & 0x10; // When a 1 appears in bit 4, it appears in the output. This can be extended to 16 bits, and a any bit delay you want there.

Another way is to have a timer running at a high clock rate, say 1 mhz.
You could

do
{
i=input;
TIMR=0xFF-4;
while(TIMR!=0) // OR you can use the overflow bit in the timer while(timer.overflow is false)
out=input;
This works by setting the counter time out for the timer. Assuming it is running at 1 mhz, and you want 5 cycle delay, and it counts up, subtracting 4 from ff will give the timer 5 ticks before it goes to 0

This technique can give you really fine delays if you are running the clock at a higher rate. Most timers can be set to the clockfrequency / 2, 4, 8 or 15. At an 16mhz oscillator and a divider of 2, you are ticking at 8 mhz 125 nanoseconds (24mhz is 41.6 ns) so at 8 mhz you have 124 nsec steps. There are other ways of doing it also. If you want a true binary counter, and you have 18 I/O's you can simply treat two 8 bit ports as a 16 bit counter and increment it each sample time.

[Davor]

When you need to fine tune a 10us delay, 0.6us is simply not good enough. So the first sample is not a place to try with varying delays. A GB sample delay takes longer, and therefore a much better place to play with delay variation.

[Sean_Goddard]

Change the 330K in the snubber to a 86K and lose the series resistor, then re run the sim

[satdaveuk]

Way behiind with this project

Its got to page 26 now,and can see you all have been busy bees, Have my PCB sorted since around page 7 and thats far as I got due to business and family time.
When time permits will ponder through from the beginning and work my way back here and get involved like the old days.
Happy new year to you all and god bless.
Regards
satdaveuk