.....because I am not interested in that approach.... and no immediate application/need. Actually, I want to look at EM fields and interactions up to about 100KHz. Or, within the BP of best audio amplifiers, switch supplies and mixed signal pcb. The other things would be nice bonus for the circuit sim guys around DIYAudio. Especially when going from SIM to build and then finding unstable circuit (osc). But, I'm not so interested I will spend a lot of money on EM software ..... just if there is something "reasonably complete" to learn from. I have OrCAD and will get access to EM somehow but it isnt as nice as having it all at your finger tips to explore.
THx-RNMarsh
Last edited:
Better yet, in some schools if you're a student of any kind, you can have access to this or many other software packages, even remotely from any location. You can download and licence via VPN tethering, or run the school's own computers remotely. All for free, as long as you are a registered student.
Audio ground planes
I brought up the audio ground plane issue back in June:
http://www.diyaudio.com/forums/solid-state/248105-slewmaster-cfa-vs-vfa-rumble.html#post3962665
My understanding is that the corner frequency of the mutual inductance of the ground plane and the resistance of the ground plane is the point where the return current stops following the path of least resistance.
If you think about it this way, a thick aluminum plate has a corner frequency of 60Hz and would be much more suitable for audio. A thick enough aluminum plate would cover all frequencies in the audio and RF range. How about a 1cm thick heatsink baseplate?
I brought up the audio ground plane issue back in June:
http://www.diyaudio.com/forums/solid-state/248105-slewmaster-cfa-vs-vfa-rumble.html#post3962665
My understanding is that the corner frequency of the mutual inductance of the ground plane and the resistance of the ground plane is the point where the return current stops following the path of least resistance.
If you think about it this way, a thick aluminum plate has a corner frequency of 60Hz and would be much more suitable for audio. A thick enough aluminum plate would cover all frequencies in the audio and RF range. How about a 1cm thick heatsink baseplate?
I do use gnd plane in audio/low frequency instrument design where the currents are small, the circuits are Class A so any coupling won't add to distortion
you do have to watch Class B current loops much more carefully, explicitly plan, route the return path
of course this isn't any new discovery for someone with 30 years of precision instrument design experience
you do have to watch Class B current loops much more carefully, explicitly plan, route the return path
of course this isn't any new discovery for someone with 30 years of precision instrument design experience
A quibbling amount, but I would like good behaviour up to 200kHz - a nice, round safety margin of 10 times, especially to ensure the mechanism feedback behaves as expected ...
I was in Asia that month but read it.... its what got me to bring up my experiment now in july and I am back in USA.
There is also a range of freq where some is going this way and some currents are going that way. Its all over the pcb areas. And, that is with just a single wire path.
I have a couple probes which picks up localized fields and another which picks up from a distance. I'll throw some info out. But, bottom line is, I think bettter pcb designs could be done if one understood more about the EM's of the layout.
I have to find the test probes etc I used and will then show them and a few things I learned as to magnitude of the issue. As I already showed/demostrated that skin effect is present but not an issue if the pcb trace is thick enough. But low freq magnetic fields are very hard to shield against and inside an amp is a mess of it.
My goal is only to bring it up and show some possible significance. Then if interested, one can go further.
THx-RNMarsh
Last edited:
I'm currently laying out a S/PDIF + DAC + Buffer board and am 'agonizing' over whether just to lay the ground plane across the whole board (its 4 layer) or split it into 2 sections so S/PDIF + DAC and then a separate plane under the buffer amplifier.
The DAC analog output in this latter case flows acoss the split ground plane. In this split ground plane version, there is via a separate layer ground connections going back to the central star ground that joins thetwo halves of the circuit.
I'm currently tending to the solid ground plane, but wonder what the current flows are like - there is a 50 MHz clock, the ~3 MHz S/PDIF signal and of course the 192 kHz between the receiver and the DAC . . . an interesting challenge!
The DAC analog output in this latter case flows acoss the split ground plane. In this split ground plane version, there is via a separate layer ground connections going back to the central star ground that joins thetwo halves of the circuit.
I'm currently tending to the solid ground plane, but wonder what the current flows are like - there is a 50 MHz clock, the ~3 MHz S/PDIF signal and of course the 192 kHz between the receiver and the DAC . . . an interesting challenge!
I think Bruno had something useful to say about that in his G-word article. I believe he came to the conclusion that the plane should NOT be split.
Jan
Tom Van Doren had some good slides on this as well. In addition, the same methodology applies to wires outside a closed box, such as IC's and safety bonds.
Current takes the path of least reactance.
Actually, no.
I had figured that by pulling the cascode emitter with 100 mA (for example), the cascode's collector resistor drop would keep the cascode's collector well above the base level, so it would never saturate, then the 100 ohmer to the rail would be used to pull the fet off.
Yah, I figured smaller was faster, I was concerned with total gate charge. As to minimum inductance, I do have some ideas on how to do that, it'll be fun to try. Theory, idea's, build design, hardware build and encapsulation, and first tests I'll document well for the guys here as well as the researchers who's papers I've been reviewing.. At the moment, I'm starting to cut metal for the rotating chuck of the plunge setup. I'm still trying to figure out how to measure runout once it's below 250 micro-inches. That's about the limit for me at this time..
1. The latest round of papers I've reviewed use 20 volts and about 1 to 2 amps, the current controlled by a series resistor, the 10 ohmer in my schematic.. The kerf of the cut depends on the voltage and the dielectric medium's breakdown voltage. The problems I've found with the latest rounds of papers is that they do not control the setup inductance which kills risetimes at that level. Many conclusions are reached regarding onset of spark, rate of current rise, voltage, and resultant cut rate and surface finish, but they are based on terrible waveforms and measurements. They tent to keep the feed wires out of the liquid, so by design create a large inductance loop AND drive a .001 inch diameter wire with current for inches outside the liquid (given the process has a variable spark rate, this can excite mechanical resonances in the wire in the air, messing up surface finish). The process is typically designed as a closed loop system, with the feed rate (cutting rate) of the wire dependent on the average current of the spark. 150 kHz, 30 nSec wide pulses, and 1 to 2 amps was in the last paper, and 1.5 micron per second feed rates as a maximum.
They hit a speed wall, and I suspect much of it is due to the lack of control of inductance, with debris clearing a far second effect.
My design will have the switchfet, both resistors, and capacitive storage all put into an encapsulated structure that will be under the work in the dielectric fluid. I'm looking to decrease the kerf width via lower voltage while retaining a cutting rate that doesn't wait for the final collapse of the universe to finish, and get really good surface finishes.
Carbon electrodes seem to work the best with respect to electrode wear. My problem is the initial app would need an electrode about .003 to .004 inches diameter, I suspect that I may just have to use drills of that size, they are easily purchased and not very expensive.
ps..demian, thank you very much for your hints and comments.
jn
Last edited:
JN,
I'm not sure of your application for the EDM machining but suspect you are trying to make watch or clock gears? I never looked into the technology myself as I always had others do this type of work when making aluminum molds but this has been done for so long I am wondering why it is so hard to find the information you are looking for. From what I remember the electrodes were made from carbon blocks and machined to shape. The most common reason to use this tech was to create sharp corners or square cutouts that just can't be done with a rotating cutter of any kind. I would think that any patents for this kind of machining would be long since expired and would think that some of this information must be available from the machined molds industry, have you looked anywhere like the Society of the Plastics Industry or one of the other plastics mold making groups?
ps. I know you are on LinkedIn and perhaps one of the mold making groups there would be a good source of information.
I'm not sure of your application for the EDM machining but suspect you are trying to make watch or clock gears? I never looked into the technology myself as I always had others do this type of work when making aluminum molds but this has been done for so long I am wondering why it is so hard to find the information you are looking for. From what I remember the electrodes were made from carbon blocks and machined to shape. The most common reason to use this tech was to create sharp corners or square cutouts that just can't be done with a rotating cutter of any kind. I would think that any patents for this kind of machining would be long since expired and would think that some of this information must be available from the machined molds industry, have you looked anywhere like the Society of the Plastics Industry or one of the other plastics mold making groups?
ps. I know you are on LinkedIn and perhaps one of the mold making groups there would be a good source of information.
Last edited:
The first problem is googling EDM. While I like electronic dance music, it doesn't help me much...
Some of our vendors used edm for the half meter thick laminated steel yokes we needed, but there they were concerned first with rapid removal, then a final cut to tolerance, but using 10 to 15 mil wire and hundreds of volts.
Micro-wire EDM is a more precise use, and the lit out there is rather limited.
My first app is plunge hole cutting on watch balance wheel staffs for repairs because a friend needs it, but yes, I'm looking further down into EDM for watch and clock gears. My first app there will be EDM of carbide tool bits to make the cutters for custom gear teeth. I love the fact that EDM goes through any metal known..
Eventually, once I'm further along, I'll visit the on site EDM facility and get some process info from them.
jn
Ya,
EDM dance music isn't going to cut it for machining your gears and shafts! Can't stand that noise myself anyway! I'll see if I can find you some information from one of my plastics groups, worth a try anyway. I know they never used EDM machining on those original watches and clocks I have so wonder why you are requiring this level of machining for such an old and well known art. I haven't done any watch repair since I was really young but have a grandfather clock and a few wall hanging clocks that need repair. Some clocks I have no problems looking into but the grandfather clock with all the chimes and different tones you can select and multiple hanging weights is something I have never messed with. Could be interesting or a cluster F..k.
EDM dance music isn't going to cut it for machining your gears and shafts! Can't stand that noise myself anyway! I'll see if I can find you some information from one of my plastics groups, worth a try anyway. I know they never used EDM machining on those original watches and clocks I have so wonder why you are requiring this level of machining for such an old and well known art. I haven't done any watch repair since I was really young but have a grandfather clock and a few wall hanging clocks that need repair. Some clocks I have no problems looking into but the grandfather clock with all the chimes and different tones you can select and multiple hanging weights is something I have never messed with. Could be interesting or a cluster F..k.
Thank you. I suspect they don't do wire, and I think feature sizes in the single digit thousanths is not a plastic forte'.
Normal technique to fix a broken pivot is drill a hole in the end and loctite a hardened pin. Easily done as long as the pin is .020 inch and up. A watch balance staff requires something closer to .004 inches hole and pin to do the same repair. Otherwise, one has to machine a new balance staff and then harden it. This procedure is becoming less and less available across the USA as the talent is dying off, and taking 3 to 4 hours to make a new one takes the repair up in cost at 150 bucks per hour.. EDM'ing a hole that size and laser welding a new pin in is one method I'm working on for my friend. The EDM has to be fire and forget, but machine time estimates at 50 dollars per hour. If I can get the accuracy and reasonable speed, my friend will probably advert in NWACC and get some business. If it doesn't work, I still benefit in that I have learned something new. If that something new also helps advance SOTA, so much the better.
Yes, the strike and chime clocks certainly look daunting at first blush. Knowing what pictures to take during dis-assembly is a skill the beginners tend to learn very very quickly.
jn
Last edited:
Either way can be better in a particular case, but much depends upon the layout, especially for analog and mixed-signal boards.
The better the layout, the more direct the current paths will be, and the more the return currents will remain near them.
Now I better understand what you are doing. EDM a hole in the end of a balance shaft etc. I also thought you were using it to cut gears. I was going to suggest photomachining for the gears as an alternative but this is different. Everything is smaller making it more possible to use higher frequency stuff.
JN,
Thanks for the clarification of what you are trying to do. I know that wire EDM is rather common in machining operations, at the same time your size requirement seems to be the issue. That would be one small wire size you are talking about and I wouldn't think you could have much length to that without having problems with alignment and support. Not sure what type of wire you would be using but think perhaps tungsten is what you would be looking at. I'll see what I can find out.
Yes that grandfather clock is amazing, multiple chime selection and multiple weights with plenty of tubes for the sound. It wasn't a cheap clock so worth repairing. I think the oldest clock in the house is a water clock, hasn't been working for years, didn't want to mess with the authenticity of it by repairing it. You add water and as the water level goes down it keeps time. I live in a house full of antique clocks. I still have my original Pulsar digital watch somewhere, I think it was the first digital watch commercialized with its red crystal face. I still wish my old Seiko chronograph with self wind was still working correctly, one of the best watches I have ever owned, kept perfect time.
Thanks for the clarification of what you are trying to do. I know that wire EDM is rather common in machining operations, at the same time your size requirement seems to be the issue. That would be one small wire size you are talking about and I wouldn't think you could have much length to that without having problems with alignment and support. Not sure what type of wire you would be using but think perhaps tungsten is what you would be looking at. I'll see what I can find out.
Yes that grandfather clock is amazing, multiple chime selection and multiple weights with plenty of tubes for the sound. It wasn't a cheap clock so worth repairing. I think the oldest clock in the house is a water clock, hasn't been working for years, didn't want to mess with the authenticity of it by repairing it. You add water and as the water level goes down it keeps time. I live in a house full of antique clocks. I still have my original Pulsar digital watch somewhere, I think it was the first digital watch commercialized with its red crystal face. I still wish my old Seiko chronograph with self wind was still working correctly, one of the best watches I have ever owned, kept perfect time.
- Status
- Not open for further replies.