piezo ceramic project

[hoola]

I placed a small square of heavy felt between the stack and the scale and lost all indications of thrust.  I then placed in thin felt and finally a thin slice of dense foam rubber with similar results. This seems to show that any readings were due to physical vibrations reaching the scale. The attempt to eliminate this issue was with the long lever arm and the thin thread with a small tension onto the scale, and that is apparently insufficient for that purpose, as this latest failure seems to point out that vibrations did follow through to the scale, the thread acting as a conduit of these vibrations instead of a suppressor,  which the scale misread  as weight change. 

the edit button is gone for some reason, and I did want to add that obviously, the felt or foam rubber absorbed the vibrations. 

[hoola]

It seems that if any thrust is developed in the system, it would be very difficult to measure with a scale, as vibrations or other issues would swamp that signal. The plan of sampling the output of each piezo to drive 5 horizontal output tubes which drive a matching stack at much higher voltages is next. I have already begun assembling components to do so. Proof of thrust would be best shown by doing actual work, that is physical movement in a pan balance type setup, which is the tentative plan.   A spinoff of this project is the scan traces which after a year of tests, are still delivering novel and interesting shapes. I wish to capture the traces straight from the input to the scope, as still pictures do not show the details well, and do a screen print onto fabrics such as Tee shirts or other clothing articles. I am not sure if the art aspect of this would be acceptable on the amateur science forum, in the other sciences forum. I do need to find out if a scope signal can be translated directly to a stencil or to photocopier type paper, which I understand is used to do iron-on transfers onto cloth. I have found that simple photos can be copied onto special copy paper for iron-ons, which I plan on doing first, but a direct transfer from signal would be preferred for clarity. If the mods would please rule if I can open another thread here dedicated to that end, or on the other sciences forum, I would appreciate your attention.

[hoola]

The false thrust through vibrations means that the prior establishment of parameters used are very likely irrelevant to the parameters of a real thrust signal.   A new series of parameter setting scans are being started to see if any stay effects might show up to offer a starting point for the tube driven arrangement.  I have glued a 2mm thick pad of foam rubber onto the end of the stack thinking that will block vibrations more effectively than felt.

[hoola]

further tests of the suspended stack offer no alternate frequencies of interest, so the usual frequencies of 115, 125 and 161khz are to be used initially in the tube driven arrangement. It does seem that any true thrust, or a "DC" component of the overall signal, is buried within the "AC" inevitable noise of the system, which is what the scale misread as weight change. The question of how to transfer a useful signal from the stack to the tubes without the drive signal, which is of a large P-P voltage, from overwhelming the "echo" of this stimulation that occurs during the delay time, which is accordingly of much lower voltages and is the only signal I want transferred onto the tubes.  I can't simply attach to the stimulation contacts of the primary piezos  to the tube inputs for this reason, so will use the edge contact arrangement I previously used to see what was happening at each active element in early experiments. I will have 5 piezos edge contacting the 5 active ones in the primary stack, and use these to pick up low level signals to be delivered to the tubes without excessive drive signal interference. These five "sensor piezos" outputs will be passed on to the tubes, to amplify this signal to the center five of the new, or secondary stack, with 2 end "buffer" or turnaround piezos added to mirror the primary stack arrangement for a total of 7 piezos. The secondary stack will be mounted on a pan balance arrangement to see if any destabizing of the balance can be achieved with switching the generator on and off. The primary stack will then no longer be attached to a scale, only to be used as signal source. I plan on using 6BG6 horizontal output tubes, used in televisions from the 50s. They have a P-P plate capacity of 6.6KV positive and 1.5kv negative. Since the circuit is low inductance, there should be no excessive "flyback" effect to concern about.

[hoola]

While the levels between drive and echo signals may be made more equal with the previous "edge contact" idea , it seems a more active arrangement will be necessary to achieve the desired equal values. The overall idea is to replicate the waveform present in the secondary stack that appears in the primary one, only much larger.  If I have a signal direct from the primary stack going through a series of step down resistors, lowering the max voltage to the tube grids to some voltage driving the tubes appropriately, the echo signal will be too low to do much for a tube output in this direct path. The idea of also passing the signal through a parallel signal path with a buffer transistor that can be switched on during the echo phase, then shut off, or muted, during the drive phase seems attractive. The buffer will have to be a emitter follower circuit so that the phases of the direct and secondary paths will be in accordance when combined at the tube grids.  The use of pin 10 to drive the delay circuit has worked well, and it could be used again to forward bias the buffer transistor in echo mode, which would be normally held in the off mode. When the delay circuit has completed, the scan begins at pin 1, and this could be sampled over to another transistor, which will act as a mute for the buffer, pulling it's bias to zero, killing the signal through it until the next scan is completed, and pin 10 is once more activated, turning the buffer back on. In effect, the pin 10 will turn on the buffer to allow signal pass in echo mode, then stay on until pin 1 activates the mute transistor, killing the signal once more  through the buffer during scan mode. Since emitter follower circuit gain is necessarily less than unity, some stage of gain may be required post matrix of the direct and buffered signal paths, if the matrix doesn't provide enough voltage drive to the tubes, which would be easy to do. The delay is sometime changed, and regardless of duration of the delay, the activate/mute arrangement of the buffer will make the new arrangement agile in it's response to any changes made in these delays. Changing the delay time is one of the parameter changes that alters the waveforms to provide either positive or negative thrust signatures.

[hoola]

I have read lately that superconducting material may not only exclude magnetic fields, but perhaps also gravitational waves. Could the piezos I am using show any altered effects at liquid nitrogen temps?  I doubt the piezos would be superconductive even at ultra cold temps if tested. Liquid nitro is cheap and readily available so I will try it when the next arrangement is functional and waveform patterns have been established just for fun.  Hopefully they won't shatter when run cold.  With the new setup I should be able to pump 20 watts into each individual element.  Up to now, only about a few watts have been used per piezo. The parts for the high voltage series are being assembled, and I hope to have things underway by June.

[hoola]

Since a problem of shock waves travelling back in the stack and mostly neutralizing any potential forward thrust is a main consideration, what if the stack was circular...?..then, a thrust bearing shock wave might impart a rotational force to the wheel assembly.  Perhaps some 3D printed wedges to go between them could offer a reasonable transfer of energy between them, and to provide the overall circular shape. The bigger problem is to induce the shock wave to have a preferred direction of travel, so a sub system would have to be developed that "neutralizes on the fly"  using the several piezos just aft of the element under stimulation. I am using timing intervals to attempt the neutralization of this return shock wave in the linear stacks and see some evidence of that happening. Most everything has been assembled to the next set of tests, and the tubes are on order. Hopefully in two weeks things will be underway.

Edited July 1, 2021 by hoola

[hoola]

It does seem that with the circular stack idea, by simply removing one element and leaving a small gap in the circle , that would offer the same basic system as with the linear stack idea, with a starting and ending element. This arrangement would offer a more practical way of getting rotary motion without any added complications of a revised NFB circuit that I tried early on with the linear stack, which did seem to offer some thrust signatures, but not as much as with the timing delay format.

Edited July 2, 2021 by hoola

[hoola]

 I was hooking up the original stack with my new one,  wiring in parallel between them and positioning them one up, one down, each located at opposite ends of a 25" balance beam. While soldering them together to hook up to a common feed from the driver chassis I realized I had made a simple wiring error, and as I was unsoldering a drive wire from piezo #1 (which was an error, as piezo #1 and #7 are turnarounds and used as signal source for the scope, and doubling as "turnarounds" for both). I  then hooked the wire onto the piezo #2tab (the first drive piezo), and as I was withdrawing my pencil from the tab, an arc occurred, accompanied with a rather loud snap. This was very surprising as the arm was on the bench, unhooked from the electronics, and away from anything I can imagine could have caused such a high induced voltage. The 15 watt soldering pencil was the only energized thing near. I did not bump, jar or hit accidently the stacks. I went ahead and continued to hook up the parallel harness, as I was only half done at that point, and when I moved the tip away from the piezo#3 tab, another arc occurred, and this time I happened to notice the arc was between the inputs of piezos #2 and #3. Upon completion of the task, the stack was installed and energized, as I wanted to see how the would interact powered in parrallel. The thrust signatures on the scope and appeared unchanged with the new arrangement, but no movement of the pan balance of course and as expected.  I would not have written this but for the two arcs that occurred. I did see arcing on the piezos surfaces as I was modifying the tab hookups with nickle paint. As the paint dried with a blowdryer, tiny fuzzy multiple arcs did occur, but no sounds, that being about a year ago. I have the 5 drive tubes mounted and wired, and attendant supply done to drive the tube grids with a -90 volt bias source, a +340 supply for the screens, and of course a 6.3volt ac for the filaments.  I have finished a separate 5KV, 60ma supply for the plate drives. The supply can be easily changed over to 10KV if needed later on. In either case, a variac will allow a gradual smooth increase in voltages.  I have the 1/4" polycarbonate sheeting, and will begin fabbing up the low pressure helium containment (noise abatement) for the new stack today and should have other detalis worked out soon for a proper test by the weekend.

 

[hoola]

the piezos are 5.5mm thick, with a 1mm poly washer and a .016 sheet copper between them, so the arcs were a little over 1/4 long, occurring between the tabs, which are bump outs on the circular copper sheets in order to solder on the drive wires. I have not been able to reproduce the phenomena so far with any cautious physical stimuli such as tapping individual piezos or bumping the central rod. I do live in a desert area, so low humidity could have been a contributing factor by allowing charge to accumulate.

Edited August 22, 2021 by hoola

[hoola]

I failed to mention the method of my soldering, which is "ball soldering" and I leave the pencil on about a half second longer than needed and add the little extra solder, then flick the pencil away while the tip scrapes off the connection. In this way a roundish blob is pulled off the connection and then has enough heat to snap back from surface tension, leaving a smooth shiny round surface. Could the act of the quick and sliding removal trigger charge? Or the rapid drop in temperature, or a change in shape of the drop as it congeals and forms a minimum surface? Both sparks occurred exactly as I pulled off the pencil tip. Since this oddity took place twice in quick succession, it seems worthy of checking out further, as the sparks seemed substantial enough to have knocked me out of my chair, but mostly from the unexpectedness of the situation. I will go ahead and hook up a fet voltmeter to the suspected tabs and see if I can replicate the sparking by resoldering the tabs once more.

[hoola]

the pyro electric effect of perovskite material is used to sense temperature change, and does emit a charge when stimulated, but seems very unlikely that the change in heat was sufficient to make any significant charge sufficient for arcing, and the heat was almost entirely applied to the tab itself, taking care not to heat the piezo and especially the poly washer any more than necessary, which melts easily being made from common polyethelene. I did make a few attempts to stimulate charge through heating and resoldering some of the tabs, but found no charge, and noticed  one of the poly washers was getting slightly melted with the repeated heating. I planned on only having to heat them once, and did not expect to be sidetracked with this issue, so will return to it next month  when I will build another stack specifically to test for excessive pyro electric charge, and proceed with getting the planned tests in order.

[hoola]

An excellent lecture on perovskite materials is on youtube "The Woderous World of Perovskites - with Mike Glazer". From 2017,  covering it's history and up to the current research.

[hoola]

sorry, I misspelled wonderous.

[hoola]

today I set the value for the bias to -30V to give approx 35ma idle current per tube. The plate and screen grids of each tube are tied together at approx +350V, as the 5KV supply is not hooked up yet. The cathodes are all directly hooked to ground. No AC signal is as of yet been applied to the grids.

Edited August 30, 2021 by hoola

[hoola]

I have tentatively figured out the melted chocolate incident. I must have placed the chocolate on the generator where it melted, then moved it as I began doing the first of the 600V experiments over to a shelf against the wall. and away from the generator, although I don't remember moving it, thinking it had always been on the shelf.  I have thought this as a possibility as the generator's power switch is on the secondary of the standard 60hz transformer, leaving the primary energized if the unit is plugged in, and does so to keep the temperature inside the unit warm, and therefore the top of the case where the chocolate container would have been. This is done for more stability when switched on. Sorry, I should have come to this conclusion and posted sooner.

Edited August 31, 2021 by hoola

[hoola]

The vacuum chamber is approx 80 cu. in. interior volume, and nearing completion. I have suspended the stack within it on a coil spring and so giving it more degrees of freedom of movement, not being fixed directly to the arm, now free to move  in the vertical Y directions, due to the chamber being mounted on the arm, but also in limited amount in the X and Z directions within the chamber itself, and even to rotate if any forces should allow it to. I don't think I will need to completely evacuate the chamber, and have decided not to get helium for now and to see how a partial normal vacuum will keep the noise down to a reasonable level.

[hoola]

Today's testing shows that the 1/4" polycarbonate is in itself very sound deadening, and may not need to be even partially evacuated during operation. The enclosure should preclude any false thrust due to acoustic pressure from the stack, as any should be neutralized, staying within the containment. 

[hoola]

When you switch scan directions there is little to no correlations to the scope display simply being inverted, as one would think, as they are laid out with symmetrical spacers, washers, etc.,  instead they are almost entirely unrelated in appearances with occasional  rare exceptions. I have yet to see a clear reason why this should be he case. Asymmetries in the torque within the stack seems responsible for some of it. These tests are done with all piezos on.

Edited September 6, 2021 by hoola

[hoola]

Today I worked on the original stack, replacing the copper contact pads as the wire hookup tabs were beginning to fall off due to a year of constantly soldering and unsoldering them. I replaced the pads with the .006 copper as with the new stack, in a further attempt to equalized their responses when working in tandem. Tests done after completing today's repairs showed a closer correlation in the two displays when scan order is reversed.

[hoola]

the next test will be with the "edge contact" method used to drive the tubes, as they do offer high isolation from the initial pulse. A  properly designed electronic selective gate system could have a near complete isolation if desired,  with the added benefit of having a control of relative values of the two signal strengths.

[hoola]

today I hooked up 5 piezos in the edge contact method and got good signal output on all five. I placed one of the signals to the grid of output tube #1 using a supply voltage of 500V dropped through to a load resistor of 20K@25watts to the plate.  The DC level of the plate  measured +225V, so the AC signal at the plate is in the order of half that,  or around 100V P-P. although this is an estimate, as the scope is rather dated and probably inaccurate, but a good outcome of this initial trial to drive the tubes in this fashion. The 20K resistor got hot, but not too hot to touch, so is in the ballpark of an appropriate load resistor, but is probably on the low side. The tube current is 10ma when not supplied with B+,  representing the screen current. When the B+ was applied the total tube current increased to 20ma, well within the max ratings of the 6BG6 glass tubes. I have not hooked up this  plate signal to the new stack yet, only have measured it's output, but will shortly have all five tubes in operation and driving the new stack. The tube I have tried first is an old RCA and probably 50-60 years old and checks rather weak on the B&K 707 tube tester, but not leaky or gassy. My intention is to replace these with NOS tubes when I have worked out all the details with the used ones.

Edited September 14, 2021 by hoola

[hoola]

Today I removed the matrix from the two sensor/turnaround elements  of the original stack and looked at their outputs separately and can see more clearly the coorelation between maximizing assymetries in one element vs. minimizing assymetries in the other  that sometimes occurs. A typical display seems to divide itself in half across a horizontal line, offering a "virtual zero reference line" and the two halves can be made to meet and form interesting relationships with the other as they do so, but only as they approach closely, as if they do meet, they tend to collapse into random noise, indicating a negation of any thrust signal possibility. A few distinct drive frequencies markedly, 115khz, 125khz and 161khz  have the property of enhancing either a positive or negative overall assymetry, indicating a potential directional trust in one sensor element, coupled with a symmetry in the other, indicating a negation of phononic energy, and so minimizing the unwanted counter thrust to any potential  thrust. 

Edited September 21, 2021 by hoola

[hoola]

The edge contact drive now delivers an accurate inverted output from the tubes plates after determining the correct grid bias setting, but at a low level that requires adding a gain stage for each tube input to deliver a useful output. Conventional transistors can be used for a 5 single channel amp board to be added onto the main driver chassis using it's existing power supply. This requirement was not unexpected as the tubes are low gain.

[hoola]

today I finished the edge contact - tube interface module and which now provides a high P-P tube plate signal, using a DC plate supply of 500V. The waveforms are clean and relatively undistorted and show the signal in it's pure state, as the secondary stack is not hooked up yet, only 90K load resistors from the high voltage supply functioning as dummy load. The bias is still in need of some refinement, as the bias setting is altered somewhat when driven by an appropriate signal. The overall cathode current seems low at this point, showing an averaged current of 5ma per tube with this robust drive signal voltage. The next test will be with the secondary set of piezos hooked into the circut hooked directly off the plates. I have eliminated the coupling caps between them as they seem unecessary as the the piezos are good insulators, and seem to have unexpectedly high capacitive value, as shown by my twin arc surprise of last month.

Edited September 26, 2021 by hoola