piezo ceramic project

[hoola]

The secondary stack does not show the assymetry that is occurring in the original one despite changes in driven frequency or delay time to the original, or bias adjustments to the tubes. Both stacks now have separate end sensor/turnaround lines having eliminated the matrixes. The main issue is not that the two stacks are dissimilar in trace outputs, but that the driven stack shows no assymetry in opposing end traces in any situation so far, and that differential in opposite ends of the stack is what I am using at the moment to surmise a thrust is possible, not so much the details of the traces, but that they maximize on one end, with a corresponding minimum on the other, therefore causing a thrust in the direction of the minimum end. If I cannot get a desired assymetry in the the driven stack, I may add a low level output from the 4017 chip to run the tubes directly, bypassing the original stack/edge contact system. I have more piezos on order to construct another stack and test again the double stack idea, where two tube driven stacks are at opposite ends of a balance beam, hanging in opposite directions. The arrangement is nice as the stacks counterweight themselves, simplifying the physical apparatus, (and doubling any actual thrust developed) They will be out the containments for these tests. In the previous test of this manner, the stereo sound they produced  was interesting also, and sometimes developed  rather pleasant tones imbedded in the audio mix rather like a sleep machine.

[hoola]

it has occurred to me that just because I may be getting good info on the relative levels of energy of the top and bottom of the stack using the end sensors, this does not indicate the energy levels within the three middle active piezos. The top sensor might sense that the top piezo is at a maximum, but it is unknown if the next one down has reached a minimum, cancelling the top piezo's phononic energy as it transpires it on the way to the bottom. The same thing can be said about the bottom one, as the next to bottom may have a relatively high energy level, again cancelling the minimum end piezo's  energy level at stack bottom, and tending to average out the entire stack energy to a negligible level. It seems I need to have 5 scopes to independently monitor with the edge-contact outputs, which  could give me more information on how the waves are developing within the stack and their relationships. It also seems that the end sensor piezos are causing a distortion of the overall mix of waves, being used as 'turnarounds" as there needs to be a "buffer" from each end cap for the first and fifth stimulation to have a slight delay before the waves turn around at the cap boundary, as early tests without the buffers delivered a large chaotic trace accompanied by a similar shrill sound and no interesting traces. It seems end sensor replacements with a similar sized, but not of piezo material might be tried to do the turnaround function, and rely on the edge-contact system for monitoring energy levels. Perhaps aluminum dummy piezos could be tried first. Early tests showed that the edge contact system does affect somewhat the overall energy levels in the stack, but since they are not in the direct flow of pulses, it seems they would have much less effect of the five individual traces, making analysis easier.

[hoola]

as suspected, the piezos in the middle do seem to cancel out any possible thrust differentials in the overall stack. This was using the conventional settings of delay, drive frequencies, scan order and what elements are active, so upon scanning through the parameters I found there are certain settings that offer a top to bottom gradient of energy levels in the order of an overall doubling. One helpful setting was activating the middle three elements only.   For this test  I was scoping from the five edge contact sensors. I  have removed the top sensor piezo from the original stack and replaced it with two heavy steel washers with a thin rubber sheet between them to act as the required turnaround, and add damping. The bottom sensor piezo is still there, but not used on this test. 

Edited October 12, 2021 by hoola

[hoola]

today I hooked up a rotary switch onto the five edge piezo sensors to easily and quickly scope the behavior of the drive elements and confirm a definite advantage of using the middle three elements only. A parameter I listed on the previous post of scan order means scan direction.  The stack is mounted vertically, so reverse scan means bottom to top.The reverse advantage may be due to removal of the top stack sensor piezo and replacing it with the washer/rubber passive element. With the current settings the prior drive freq. of 125 khz once again seems a sweet spot. By turning off the end elements, it seems that they offer an extended turnaround delays on both stack ends, somehow offering these positive results. The first reliable and extended observance of repeated thrust indications above  2-3 milligrams was with using only the middle three elements, which was about a year ago. 

[hoola]

I have found that the stack puts out almost no signal to the scope on any interrogated piezo at 175 khz. This happens regardless of what piezo is activated  or not, or the scan direction. The noise also almost disappears to barely audible. The resonant frequency of the piezos are 44khz, and the fourth multiple is 176khz. This happens at a fairly rapid delay rate only, of which I can only guess at, as I don't yet have a way of measuring this. Can anyone advise on what  brand and model  of an interval measuring device they have used and had good results?  I see online of various manufacturers that make them, but don't see anything used on ebay.  I would prefer a stand alone device, as I see they are usually coupled with measuring other criteria also. Any help would be appreciated. thanks

Edited October 14, 2021 by hoola

[hoola]

I have removed the washer/rubber damping element and replaced it with the conventional piezo and still have a mimimum on all piezos at the 175KHZ area, but not as pronounced with the damper in place. I have received a dual trace Hitachi V212 scope with another on order, and the LBO 514 dual trace I bought for parts has vertical output transistor bias issues, but otherwise seems ok and I am in search for a schematic of it online. I am returning to the warehouse to do further tests to have the concrete slab floor for stability for milligram scale tests.  I will add a low level output on the main drive chassis so that I may temporarily disconnect the original stack's edge output to drive the tubes, and drive the tubes directly from the  main drive chassis. The interface assy was intended to run the tubes from the edge contacts so has gain adjustments that should have enough range to amp up the direct output, as it will be on a lower drive level, on the order of 10 volts as that is the supply voltage for the 4017 IC. 

[hoola]

I have begun mapping out areas of interest and have found several that show the desired gradient within the stack, a new one at 2.1mhz among others already listed. I have begun building the third stack, which will hang in tandem on opposite sides of a balance beam  along with the second stack,(since removed from the polycarbonate containment), both being  parallel driven by the tubes, which in turn are driven by the edge contact assy. attached to the primary stack.  I have abandoned the poly housing for the moment and will use ear plugs for the noise. I have ordered a ten turn 250K precision pot to adjust the delay feature, as it is very critical to adjust now, even with a fine tuning pot hooked up in series with it. The hitachi dual trace scope should allow me to determine the delay interval by sampling the 4017's pin 10, signifying the end of scan, and the lower trace to it's pin 1, signifying the start of scan. The common scan rate to both traces should allow a determination of the delay duration by the separation distance between the two blips, and so get a reading at certain delay control positions. Several positions should allow enough measurements to encode a counter chip to deliver a numerical readout of the delay duration without having to refer to control position markings. The higher scan frequencies naturally need a shorter delay to assemble useful traces, so the readout will have to be rather broadly engineered, as I guesstimate the delay for useful traces will range from 1 millisecond and down to 10 microseconds, depending on scan drive frequency.

[hoola]

the shocks from the new stack while soldering leads to the tabs has happened again today. This time the charges were minor, but enough to feel it and I was anticipating it. I suppose this phenomena is well understood, and I won't mention it again unless some aspect of it requires a followup, but I am using a 15watt pencil and the heat input has to be negligible, and is surprising to me that the process appears to be so efficient.

[hoola]

I guess the followup will be that the charge likely precipitates from the removal of heat, as the pencil could only heat up the element slowly, and due to the mass of the ring the temp quickly drops into the ring body and the copper tab material which is a good heat sink. I found out that the neon transformer I am using has a current limiter within the sealed chassis box. I had noticed that when heavily conducting the tubes, a voltage droop would occur, more than I had anticipated. The current limiter is in the central tap, so the supply can be single ended and not use the center tap, and adjust the variac to compensate. I measured the resistance of the current limited center tap, and it showed open circuit, so there is something more complex inside other than a simple limiting resistor

[hoola]

tonight I hooked up the entire tube driven assembly and had some encouraging results. I got a reliable weight increase of 3-5 milligrams with occasional readings of +7-8. This is nothing new, as I got similar results with the SS unit with it's 600V supply over a year ago, however the shortcomings of the neon transformer are becoming apparent, as if I increase the supply volts to the tube plates much over 600V,  the noise level and chattering that comes from the transformer becomes excessive. Even the piezos mimic the sound and do so even with the drive signal to them turned off. Apparently the current control circuit within the transformer is the cause, and makes that neon sign noise that I had always attributed to the neon tubes themselves.  I will rewire the power supply to remove the noise on the high voltage line and hopefully will smooth out the supply and can continue on with the high voltage tests. Sorry about the underlining, but can't get that function to shut off.

't

 

 

 

 

 

 

 

v, the transformer get noisy and that noise is reflected into the tandem piezo units even with no inputs,  causing me to remember the noise an old school neon sign made, having been around them in store windows and the like years ago. I had surmised that the noise was from the neon tubes themselves, and it seems that the transformer current regulator circuit is the cause. The transformer and piezos chatter away sounding just like a standard neon sign. A conversion of the power supply is in order to eliminate the buzz in the DC supply, and increase the current capability, 

note...disregard the second paragraph, it turned blue and disappeared after writing it, so I had to re write it and the underlining got stuck on, then the lost paragraph reappearerd when submitting the reply, but doesn't show up in the edit function, so I can't remove it.

Edited November 15, 2021 by hoola

[hoola]

I am very happy to report that due to a simple loose connection in the HV supply, the chattering noise has been removed. This seems to mean that the neon sign noise is once again the result of the tubes, or largely so and I don't have to rewire the supply. The transformer has a nice smooth hum now, so I increased the voltage output to 2.5KV for a brief period and got a stable 11 milligrams of weight increase. I can turn off the weight gain by several means. I can simply shut down the generator, alter the delay timing, drive frequency or shut down the HV supply. The question remains for now if the readings indicate a real weight change from thrust or from simple scale vibrations. Only when I do a balanced arm test that shows actual movement can that be settled. I seems likely that scale indications are partly due to simple vibrations, but if there is a real thrust appearing, it is imbedded within it, and of a certain percentage of a given reading. Part of my problem is that the scale shuts down after a few minutes automatically, and i have remove the lever arm from the scale and start over, and the scale takes about a full minute to stabilize to the tare weight of the slightly overbalanced arm, which is about 6 milligrams. This gives me a limited time to do a particular test in most cases, but sometimes the scale "forgets" to shut down, but I don't think there is a way to make it do so. I do have the scale with the wired remote on/off button mod that allows me to activate it without touching the scale itself, which is quite a help, as pressing the button on the scale is itself a major source of arm jitter.

[hoola]

I have seen that the highest useable voltage  is 3KV, when I heard the arc snap within the stack. I backed up the voltage to just below that and did more scale tests, the results were not much, if any higher than with 1.5KV. This is disappointing as a member of the woodward team was indicating that Kaluza Kline infers an interaction that increases at the 4th power of the voltage. I have yet to hook up the end piezos on the two stacks on the arm, only am relying on the readout from the primary stack, which does certainly give access to how to proceed, but I need the two matrixed end piezos of both stacks driving another dual trace scope. I will switch between  scope inputs for now, so as to do another torque test on both arm stacks and try to equalize their individual responses.

[hoola]

today I came by some vintage surplus interior 5 conductor phone wire that has tinsel wire and is extremely thin, soft and pliable. I have subbed them for the normal wire leads between the tube output connectors and the double stack assy. input terminals. The scale settles down quickly and the "creep" up and down as the original wires relaxed is almost gone. With this mod I recorded a 114 milligram weight increase using 1 KV as tube supply. This I have cut down by 50% to 57 milligrams, as the actuator arm that rests on the scale is extended from the central pivot point, and sticks out half the distance of the arm ends where the stacks are positioned, so there is some force gain by lever action. I chose the central point to mount the actuator arm as the vibrations coming down the arms hopefully will cancel out in the center. Since some vibrations must still be coming through, or even enhanced by this setup, I put in a 1/8" felt pad between the arm and the scale and read a more reasonable 6 milligram weight gain.

Edited November 16, 2021 by hoola

[hoola]

after further fine adjustment of position of the actuator arm, I placed a 2.6 gram weight on the  left stack end and read 5.2 grams on the read scale, giving a doubling of actual scale readings, hence the 50% reduction in all measures. I placed a 3/8" thick piece of felt between the actuator arm and the scale and with 1KV tube supply got 9-10 milligrams increase, and upon increasing to 2KV got 20-21 milligrams increase, a near linear doubling. If the reading indicates a predominate proportion of thrust, the felt being an effective block to vibration, the true thrust is 10 milligrams at 2KV. This is approaching the level of removing the scale, and potentially seeing arm movement.  The stack seems more sensitive to high voltages today, hearing some sizzling and frying at 2.5 KV as the humidity is up this morning. The interface module that takes the output from the edge contact assy. has individual adjustments for each output. They are all set at midpoint so as not to overdrive the tube grids, and they can each be switched on or off by the main chassis, but I have not yet tried a reduction of inputs with these controls, varying the relative ratios of them,  offering a more fine tuned approach to the control of the impulse peaks and valleys within the stacks.

Edited November 16, 2021 by hoola

[hoola]

I set the HV to 2.5KV , removed the scale and balanced the arm to where it was barely offset to the right side and noticed that the gap between the bottom of the right stack and the wedge I inserted under it was slightly affected when activating the stacks, seeming to raise it up just to the level of barely perceiving a change in the gap, perhaps 1/32" I then turned the main generator on and off, syncing up with the gap appearing to separate from the wedge, and upon doing so, it became obvious that the gap was increasing with each pulse of signal from the generator of about one second in duration, to the point of raising the gap bounce to nearly 1/8", and clearly visible. This is hopefully is from actual impulse thrust, and not acoustic thrust from interactions with ambient air. A further test with a reduced air pressure, helium, or true vacuum may be needed. Hopefully further improvements will make the lift thrust so obvious that those tests may not be needed.

Edited November 16, 2021 by hoola

[hoola]

I have re - introduced the scale and adjusted the main generator up to  2mhz, and have achieved similar forces of 10 milligrams or so. I do this to see what is the highest drive frequency that is effective for generation of thrust and it seems that the 2mhz setting is near or at the top of the  useful frequency band. The advantage to using this higher frequency is that the entire apparatus in now silent, and the hum of the HV power transformer loading down by tube conduction is the most obvious way to determine stack stimulation. Next will be to determine the lowest generator setting to give thrust readings.

[hoola]

I have re - introduced the scale and adjusted the main generator up to  2mhz, and have achieved similar forces of 10 milligrams or so. I do this to see what is the highest drive frequency that is effective for generation of thrust and it seems that the 2mhz setting is near or at the top of the  useful frequency band. The advantage to using this higher frequency is that the entire apparatus in now silent, and the hum of the HV power transformer loading down by tube conduction is the most obvious way to determine stack stimulation. Next will be to determine the lowest generator setting to give thrust readings.

I don't know why the last post was doubled, it was not deliberately done by me, anyway I have found multiple peak and valley of thrust points as you go down the dial until I got down to 60hz or so. I didn't change gen range and do a more precise measure, because the noise it produces at lower frequencies is so loud the piezos may become overstressed and I quickly ended that test. I have found 2 small circular chucks missing from a piezo  some time ago, I think may have been caused by a test wherein I removed the "turnaround" element on one end. I quickly shut down the driver when a noise came out of it that seemed damaging. I am not sure that was the cause, as I disassemble each stack on occasion to inspect for damage, and the disassembly after that incident was when the damage was first seen. The hum from the neon HV tranformer has turned into an indicator of thrust peaks. This new finding is pretty exciting as the current drain to the tubes could be largely dependent upon overall stack impedance. As they "talk to each other" and form certain relationships, like a self syncronizing set of pendulums, the overall impedance goes down and thrust goes up. This seems a reliable behavior, but just noticed tonight and I expect exceptions. I am curious as to why I have not found any settings that produce negative thrust, as I had seen before. I had found that the neg. thrust peaked at about 1/3 the level of the max. positive thrust. Other curiosities include that once you perturb the particular sweet spot of seeming harmony, it is not always easy to re establish it. I have also long observed that the relationship that develops when scanning down in drive frequency to a sweet spot  is hard and sometimes impossible to re establish coming up on the drive frequency as you approach the  sweet spot freq.  I will install an ammeter into the HV supply and use that when the noise background too loud to hear the hum and maybe find current peaks  that aren't apparent with hum level. At the end of tonight's test, I didn't have to look at the scopes, just listened for the transformer hum.  I have been getting a reliable 15 milligrams in many parameter settings, but the 125khz setting continues to be the main frequency used.

Edited November 17, 2021 by hoola

[hoola]

yesterday I had written that I had achieved movement of the arm a visible amount, perhaps 1/8" at a maximum and described certain details. The posting disappeared, as they sometimes do and I didn't re type it, but went on and did the low freq. test and so let me describe it now. I removed the scale and balanced the arm until it almost floated off the table top where the right side rested. Upon scanning through the known hot spots, at around 125khz I detected what I thought was a slight lift off of that side. I had noticed that the arm has an overall resonant frequency of about 1 hz when manually perturbing it, and so I pulsed the generatior in 1 second intervals with a brief off between cycles. The stack began to lift and fall at this 1 hz rate to my astonishment. After some time and calling in my neighbor to witness this, I managed to lift the arm up at the 1/8" height continuously. I observed these events and took videos of both the initial bounce and the continuous displacement of the arm. I was so gobsmacked with this that I had some trouble believing what I saw and had some difficulties in concentration  when first blogging these facts, and when the entry disappeared simply went on with the low freq tests and tried not to over react to the headache that was increasing in intensity as I tried to focus on something else.

[hoola]

I had been having trouble with the unregulated HV supply going over 3 kv  and getting some minor stack arcing when not driving them and found that increasing the forward bias of the output tubes gives a good enough regulation to keep the HV down to 2.5 kv maximum, and providing  about 2 kv  when drawing tube current. With this adjustment the tube  power outputs are increased and today I got the arm to bounce 1/4" and can easily levitate half that. If the stack arm was perfectly balanced and was on a low friction central bearing with a very low friction commutator or rotary tranformer system it might make complete revolutions and effect rotary motion. I wonder if the ultimate speed of that rotation would be held to seventeen times the speed of sound in air ( the supposed speed of the shock waves moving through the stacks) or if it relies on a more fundamental interactions and could turn faster. I would suppose this thing would move in space, so what would constrain it's maximum theoretical speed in a vacuum? The tubes are used and check only fair, so soon I will get 5 new ones and see if the power increase would be as advantageous as today's bias adjusment was.

Edited November 18, 2021 by hoola

[hoola]

a number of multi wire miniature slip ring assemblies are available, but set up for lower voltage 50-60 hz.AC use. The issue of internal arcing may be gotten around by using two slip rings, one to carry the five drives and a second to carry the ground return. The friction coefficient is not a listed parameter, but will order them for testing.

Edited November 18, 2021 by hoola

[hoola]

today I replaced the delay and delay fine tuning controls with a single 50K ten turn contol. I set the arm back onto the scale and read the what is now the expected max of 30mg/2=15mg= 150 micronewtons at various parameter settings. The 200K frequency is now my favorite as it is nearly silent, which reduces the possibility of acoustics from the stack causing misleading readings, unless there are ultrasonics that could cause an issue. Since I have two stacks at right angles, if acoustic waves are coming out, they might cancel out, lowering the possibility that there is an issue there.

Edited November 19, 2021 by hoola

[hoola]

this evening I hooked the sweep generator back up to drive the mechanism and found no useful increase in output, but the sweeping sound does mimic the "flying saucer" warble noise used in science fiction movies from the 50's, but that generator is difficult to use and the main generator was quickly subbed back in.

[hoola]

today I disconnected the driver stack/edge contact assy. from the interface module and ran a new set of leads from the bases of the output transistors and fed those directly into the interface module and got weak weight changes of less than 5-6 milligrams. The signal at the transistors bases was insufficient to drive the arm stacks properly, even with adjusting the interface gain controls to max which is puzzling as I thought i had plenty of headroom to compensate for the differential. Next  I will simply remove the output transistors which will  remove the signal drain from the base currents which will give an increase. I am trying to simplify the apparatus, and if I eliminate the drive stack/interface piezos, that will be a 50% reduction in number of stacks,  plus no need for output transistors and their attendant 600V power supply. This will be important for station keeping, as I presume there will be 3 separate thrusters, one on each axis

[hoola]

removal of the output transistor has increased the drive to the interface module to bring back the 30/2=15=150 micronewton expectation.   From now on I will designate just the micronewton calculations as I feel 99% confident that there is a valid force being developed. The 200K drive frequency is near silent, with only the hum of the HV transformer as the predominate sound. I tried hooking up an ammeter in the ground return of the HV supply and oddly got no reading., perhaps due to some interaction with the internal current limiter.  I am sure that issue is not important, but the sag on the HV will be an issue when the new tubes come in and the current requirements go up. The built in current limiter in the neon transformer seems to be in the center return line, so next I will rewire the supply for single end and remove use of the center tap. This will allow a  maximum output of 10KV, but the practical max has been around 3KV. The humidity is down today and there was little stack sizzle and no arcing at 3.5 KV. I will research what other octal base transmitter tubes with higher power could be substituted in without much circuit rewire. At this point the 4017 IC driving the interface module, driving the arm stacks are the only active elements. The scopes are currently not hooked up, as they were sampling the edge contact assy. The end piezos on both stacks have not been hooked up yet, only used for the "turnaround" function, but now I will transfer the 4 scope lines over to the end piezos on the stack arms. The end stack sensors gave a much more interesting visual display, almost a character generator function, that was not present with the edge contact pics.

Edited November 20, 2021 by hoola

[hoola]

the first line of the previous post should have read...removal of output transistors...as they are all out. Today I will break down the level arm assy. to bring the balance to neutral, as the pivot point is below the arm and will fall to one side or the other if placed in the center. In this manner I will be able to place calibrated weights on the arm and test without fear of scale interactions, and get more precise estimations. The scale will be temporarily placed back on occasion to peak out bias adjustments and other criteria.