Go Down

Topic: reading voltage of an N channel mosfet (Read 4881 times) previous topic - next topic

allanhurst

Quote
The iso made the fluid very clowdy
Excellent -that means it's doing the job of removing tiny particles.

Filter?

And if you want to go to 24v - add a 12 zener between the gate of the mosfet and gnd....

Allan

mtraven

Excellent -that means it's doing the job of removing tiny particles.

Filter?

And if you want to go to 24v - add a 12 zener between the gate of the mosfet and gnd....

Allan
sorry I was unclear, the cloudiness was caused by just mixing the iso with the oil...not by particulate.  I saved the mixture, its worth trying again.

I have already made the 24v changes, seems to be a bit better, but there are a lot of ways this can be tuned, makes it time consuming to find the sweet spot.  I'll add the zener now that you mention it, but I would like to know why / what that does.

and a filter is certainly in the cards for the future, but I need to know for sure what DE I will be using, and that remains TBD.

mtraven

still having trouble with the control software...taking a little different approach, rather than averaging ADC readings, I am basically converting each value to a true or false & then I can compare the number of non zero readings to the total number of pulses and move the electrode according to that.  It has the benefit of storing a lot more data so I can see whats going on (kind of a crude scope).  I while back I used some hall sensors (which are themselves analog) but with breakout boards that made them digital.  I think they used a Schmidt trigger or something like that....my question is, how challenging would it be to   make something like that breakout board, ideally with a pot or two to adjust the threshold and output a digital signal back to the arduino?

allanhurst

If you want to know the actual current flow just monitor the source of the mosfet. 2.5 volts == 1A

Allan

mtraven

If you want to know the actual current flow just monitor the source of the mosfet. 2.5 volts == 1A
not sure why you mentioned that, but its good to know.


lot of trials last night: 

veg oil and iso was the same as previous, no arc on approach, directly to short

ATF fluid:   this was promising, it arced about as well or a bit better than the veg oil

kerosene -- it did marginally better than oils, but not good enough to warrant using it and it seized my pump after just  a minutes, and it makes damn mess

Tried thinning the veg oil with kerosene--I can get it to pump but its no better as a DE.

In general, the electrode just seems "sticky".  What I mean by that is that it goes from an arc to shorted nearly instantly & then has to recover.  This happens even if I am controlling it by hand. In fact, sometimes I can get it to arc for a few movements and then with no change in position, it shorts.  I have swept through a range of currents, thresholds & pulse lengths with no significant changes.

I have a few more trials:  I epoxied and electrode into its collar last night to see if movement at the electrode might be the problem, will test tonight.  If i can find a suitable pump, I will give distilled h20 a try.

Beyond that, its getting pretty close to terminating the project, its just sucked way to much of time.

allanhurst

#110
May 26, 2018, 12:41 am Last Edit: May 26, 2018, 12:59 am by allanhurst
The dielectric breakdown of pure water is at about 70MV/m.

See https://en.wikipedia.org/wiki/Dielectric_strength. Other dielectrics are mentioned.

If you've got a 12V  supply, that means the gap for breakdown is 12/ 70e6 metres, or about 0.17 microns.

How big is the smallest step on your drive? - it should be a good deal  less than this or a single step will cause actual contact rather than dielectric breakdown - ie you're crashing the electrode.

Allan
ps the zener is because the absolute max Vgs of the MOSFET is +/- 20V.  Just to be safe.

mtraven

well this thing has worked just well enough to keep me from tossing it out the window....

-changing to the epoxied electrode made an immediate difference, so I do think the electrode was dancing around on me a bit.  This marked the first time I had a sustained burn, rather than just periodic sparks.  So  I spent the last few days making a collet chuck to hold the wire, turned out nice and works about the same as the epoxied one, but now I can advance /trim the electrode as needed.

-currently, it has periods of brilliance where it sizzles like bacon & sparks like crazy.... but then it will just stop working(comes down, sparks a few times & then retracts on a short)...and eventually came back....over 4 hours, it did this 5 or 6 times with no changes to any part of the system.  This is very frustrating and I have no idea what to make of it.

-as for the positioning--lead screw is 1/32" / rev (0.79375mm/rev) with 7608 steps / rev...my math gives 1.04x10^-8 meters.  That seems sufficiently precise don't ya think?  Oh and I am usuing 24v, in veg oil (filtered to 20 microns)


-Last night I started to compare this to arc welding, led my mind down an interesting path....first of all, am I trying to create an electrical arc or just sustained sparks?  Are these even different (I think no)?  So when you start an arc with a welder, the conditions at the start are not the same as while the arc is running.  Generally the electrode must be much closer to start and then is retracted to a 'run distance.'  Dependent on material,  arc welders have all sorts of start sequences (HF start for aluminum for example).  So this leads me to wonder if I need my own start sequence, thoughts?


-I attached a log of how its running, the 3 numbers are different ways of tracking the adc values, the first is the moving average--thats the only one controlling anything, the other two are just observation at this point.  col 2 & 3 are percentages, how many hits(non-zero adc reads) / # pulses....when the second column is in the high 90's, but not 100% it is arcing exactly how I want.

allanhurst

#112
May 30, 2018, 03:10 pm Last Edit: May 31, 2018, 05:10 am by allanhurst
Sounds like you have enough resolution on the drive.

An arc is a continuous flow of current through a plasma once struck. EDM isn't. Your start should be down to delectric breakdown at a small enough gap.

It might be worth increasing the inductor.

I'd still be worried about particulates and the viscosity of your fluid - clearing the gap should be efficient.

Some old stick welding systems used the current in the feed to retract the gap slightly This made starting easier - it was done magnetically. So it worked a bit like an old-fashioned doorbell giving a vibrating feed..... easier than ultrasonics?

edit:  and could help in removing eroded particles  /edit

Allan

mtraven

so EDM is not an electrical arc, good to know.

But I have news...you asking about the precision brought back an old issue that I had forgotten about.  The stepstick I am using to drive the stepper has 3 pins that control the step mode (full, half, 1/4, 1/8, 1/16)....Pull all 3 pins high and you get 1/16.  When I originally hooked it up, it took twice as many steps for 1 rev as my math had suggested...I chalked it up to bad arithmetic and didn't think much more of it.  As it turns out, there is an undocumented 1/32 stepping...I suspect it  is undocumented because its not usable (although it really seems like it is).  But more importantly, I must have had a bad connection on one of the selector pins and I suspect it was intermittently switching between 1/16 and 1/32.  So I have fixed that and it is a completely different machine now. I have not gotten it to burn as well as the best it did before, but I am clearly in control of it.  Previously, there seemed to be no reason for it behavior, now there is and its just a matter of sweeping all the variable to tune it.  Current is one of the biggest factors, having that current regulation part of the circuit is very useful.  I might like to make a few changes to it, that I was hoping you could advise me on:

1. suppose I wanted to up the max current, how might I do that?  For the record, I have been running at less than half throttle so far, just would like to know how to do that

2. suppose I wanted more precise control of current?

3. suppose I wanted to control the current with the arduino, could I PWM a signal to the op amp that would do the same thing as the pot does now?


- it seems that sparking occurs to varing degrees across a range of gaps roughly equal to the diameter of the electrode (in my case ~0.010") and now that I have better control, where in that range would you expect the best results?

last question, what program did you draw those circuits in?

allanhurst

#114
Jun 01, 2018, 10:16 am Last Edit: Jun 01, 2018, 11:21 am by allanhurst
Well, it seems as usual the gremlins were nibbling before you found yet another one !... let's hope it's the last.

1/ Yes - the max current is 2.5 ( the voltage of the TL431) / R4.  Change  R4 if required.

      The limit will be the PSU and the rating of M1 - current, dissipation, heatsink size. If necessary add another in parallel.


2/ the resolution  of the current  control is defined by the resolution of the pot - R2.

Accuracy :
: R4's tolerance - 5%? 
: The TL431 is 2% depending on the grade you buy.
: Inaccuracies in the opamp may contribute  0.1%

: temperature drift probably  < 200ppm/C  all up. 100ppm from R4, 20 from the TL431, 50 from the opamp

3/ Yes - no problem - want a circuit?

Spark current density ....  figure it out in amps/sq mm or perhaps watts/ sq mm or whatever units are used  and compare with the specialist literature - it's WELL outside my field, so I'd only be guessing.

Diagram in PSPICE, a simulation program.

Allan

ps what size inductor are you using?

mtraven

#115
Jun 03, 2018, 04:05 am Last Edit: Jun 03, 2018, 12:11 pm by mtraven
3/ Yes - no problem - want a circuit?

ps what size inductor are you using?
I would like to see how you would control current with the arduino....would you also measure it? perhaps a voltage divider reading across r4?

I have no idea the size of the inductor, I pulled a few from old power supplies and rotated them through to see how it did with different ones.  That was all before the last few gremlins, so those test will need to be re done.  The one I have in right now is 11 wraps of 16 gauge magnet wire around a 4mm iron core.  Maybe that gives you some idea, sorry I don't a meter that measure henries.

allanhurst

#116
Jun 03, 2018, 07:35 pm Last Edit: Jun 03, 2018, 07:51 pm by allanhurst
Current control :



Note it works backwards.. 255 = no current, 0 = full current.

so current = ( 255 - PWMval) / 255  x max current set by 2.5 / R4.

Allan

mtraven

thank you once again Allan!

was my description on the inductor not good enough to roughly determine size?


also I have noticed that every time I take the electrode out, the end is no longer copper colored, its grey like the steel I am burning.  I am sure this negatively effects the spark, but I am not sure why it happens....first thought was too much current, but I ran it way down and still got a grey tipped electrode?

allanhurst

Your inductor description isn't isn't very helpful. Probably a few 10s of uH.

Try increasing it - say 50 turns of 22swg - 16 gauge is over the top for an amp.

Greyness - dunno.  Eroded stuff sticking? reaction with the dielectric? Try changing it.

Burnt any deep holes yet, or just tickling the surface?

Good luck!

Allan

ps = +1 karma point for perseverance !

mtraven

Ok--so more turns, thinner wire, got it.  I have a bunch of magnet wire around, so I will probably just wind one myself.  can I use steel as the core, or does it need to be that special silicon iron?

I found an oil fuel pump I forgot I had last night, so I may try to rig it up to run kerosene again, but I am also wondering if running electrode positive might change that to.

have I made a deep hole...ummm the first piece has a bit of a crater in it, but I have started & stopped and re positioned so many times, its not really a hole.  I start a new piece last night, there is the start to a hole in it, but its probably not deeper than the electrode is wide (0.010").


Go Up