TheRadioBoard

In my further experimenting with the Kyocera part # KBT-44SB-1A Piezo Elements, I've found the following:

1) Mounted them in "Harbor Freight" Ear Protectors that often times sell
for $1.99ea.
a) These ear protectors are comfortable, and very adaptable to making
a fine set of headphones.

2) Wired them so that the right and left elements can be switched, either
in parallel or in series.

3) Tried several configurations of crystal radio sets, some with inductive (Bogen T-725) loads and some with no
internal-load from their detector stage.

4) In all cases, found that the best performance by far, was having the two elements in series.
a) This finding was concluded by the use of an, RF signal generator at set's input, and scope across the set's
audio-out (across the phones).
b) Generator was set at about 1,000Khz, at minimal Audio/RF-out. Scope was set to see the cleanest/highest
audio, from set's output. The audio is a 1khz tone, set very closely to 100% modulation-mix with RF.
c) In all tests, the sets were peaked for maximum performance, from the signal generator's output.
d) Various detector diodes were used, but of course the FO-215 was "tops".

5) Using the inductive/Bogen transformer load, found that the best performance was to uses the 5K or 10K tap, with
the elements in series. The 1.2K or 2.5K taps were the best, while the elements were in parallel.
a) The above is dependant upon the setting of the "Benny" circuit, or a 100K pot, paralleled with a .1uf capacitor.
This combination is in series with the detector, and the 40K tap on the Bogen transformer, while the black
lead is at common/ground. The average setting for the 100K pot here, is about 33K-ohms.
b) The elements in series, and connected to the 5K or 10K tap, gives the cleanest and highest audio, by far.

6) When using the Piezo elements with an external/resistive load, found that the best output in with the elements
series was with a 12K resistor, for the load. With the Piezo elements in parallel, found that a 6K resistor gave the
clearest and most audio-out.
a) Again, the above combination was far the best with the two Piezo elements in series, and a 12K resistor-Load.

Conclusion: The best combination of the Kyocera part # KBT-44SB-1A Piezo Elements, is to wire them in series and use an inductive load, such as the Bogen T-725 transformer. The next-best thing is to wire them in series, and use a 12K resistor across the phone's input for the highest and clearest audio. The resistors are not used, when using a transformer/inductive load.

I would have no problem saying that when the Piezo elements are used properly, they are next-best thing to sound-
powered headphones for sensitivity, and good/clear audio.

Hope this helps to clear-up, the usage of these elements and would welcome hearing about other's experiences that are actual/factual findings, from the standpoint of "bench-testing", in the real world of radio experimenting

Mike

_________________
I like the K-I-S-S Principle: "Keep It Sane & Simple"
http://www.peeblesoriginals.com

Thanks for the info Mike. It just goes to show that one shouldn't make assumptions. I had assumed that given their high impedance, they would be better in a parallel connection, and that's how I wired mine up. Now I guess it's back to the drawing board.

I'm going to remount mine in a pair of hearing protectors too. I had them mounted in a set of cheap walkman style phones, and that was definitely the wrong mounting style.

thanks Mike. i recently switched to a Kyocera piezo. my 1AD plate circuit ends at that piezo with a 125k resistor across it. any suggestions for improvement? i have done few Bogen tests and no Benny yet.

regards,
phil

Phil,

I played with my Acorn One last night and the 12K seemed
consistant there. However, the Bogen load is by far, tops.

I'm of the belief that the 12K resistive-load could vary
from set-to-set, and especially with tube sets.

Hope that all will keep us posted on their findings with this,
and in no-way do I feel that my initial findings are conclusive,
and nothing more than and good start.

Mike

_________________
I like the K-I-S-S Principle: "Keep It Sane & Simple"
http://www.peeblesoriginals.com

Mike,

Thanks for posting the results of your testing. It certainly helps those of us that don't have equiptment to make these sort of tests.

I have often thought about using piezo elements, and now I will get some to try out along with some of the harbor freight ear protectors.

I do have a set of sound powered phones that I made from telephone ear pieces and they seem to work ok for me, but will make the piezo set and compare them. I'll post the results.

GGus

_________________
Old age is 'when you still have
something on the ball, but you are just too tired to bounce it.'

Because of their high capacitance the impedance is actually quite low. Some are ca 3.5k 'Z' each, the poster cites 2.8k'Z' for another example. They really need a matching transformer to work well and properly in most circuits.
I played with these things a lot before getting into S/P a while back (the STM was a result of that) and today dug out the 3 piezo headsets I still have here. None of them are using the elements presently under discussion, 2 are discontinued RS piezo speakers and the other a pair of salvaged large brass disc telephone sounders. Can't say at present how their relative sensitivity works out compared to the blue Kyocera speaker.
These piezos and the old-style TELCO earpieces are the main lo-budget alternatives to S/P.

_________________
S.B.

Steve,



That's the way I see it, too.

What I was trying to doe by finding the right resistive load, as I have customers that want phones that work with crystal sets that are above the little ceramic earplugs. Most of these people do-not want to mess with a transfomer, so was trying to find the proper resistance that I could build-in, to eliminate the confusions and problems that those don't want to deal with.

Also, just wanted to pass-along that info, too.

Mike

_________________
I like the K-I-S-S Principle: "Keep It Sane & Simple"
http://www.peeblesoriginals.com

Good stuff there, Mike!
Interesting that they work better in series than parallel. Shouldn't make any diff with the xforemr but there you go. Practice trumps theory. I wonder if a ca 220k DC drain across each would change anything on that.
It's easy for some to assume that since their resistance is infinite, their impedance must also be sky-high but it just ain't so with these. The old Brush-Clevite Rochelle salt piezos were very different with their much lower effective capacitance, they'd work OK directly off an xtal set with a high value drain load resistor in parallel.

_________________
S.B.

Steve,

Actually, I've found that there is a remarkable difference in their output, when in series, rather than parallel. Have tested this over-and over, in differing situations and every time have noticed a cleaner, higher audio with the elements in series. This is especially so, while using a transformer load.

I would really like other people to run the same comparisons, to confirm my findings. What I found here was, using the signal genrator and scope to see the differences, then go-back and air-test, then saw the real differences.

Have consistant frequencies that I go to, where there are real "peanut-whistle" stations. Also, have two torches, that help me to run these on-the-air tests, to confirm what the instruments have shown.

In previous posts about these, all I saw was using them in parallel, with a 100K resistor across them. I built two like that for a customer, last week and was highly disapointed in them. Previously had noted the improvement with them in series, but found the 100K resistor to reduce their sensitivity and really ruined their audio quality, with higher-level signals.

Would like to see testing, on the bench, on the air and with instruments, from others, like I did.

Specs are great, but sure like to see more, real bench-testing, in the "real-world". Much like what you guys did to ultimise the sound-powereds, and associated matching to the real-world of circuits.

Mike

_________________
I like the K-I-S-S Principle: "Keep It Sane & Simple"
http://www.peeblesoriginals.com

Mike, if you could humor me on the static drain angle one more time, please try a 1.0 meg resistor across one only and each of both elements in series and see if it affects the performance for the worse.
I will build a conventional headset with a pair of those when I can and get back to you & the group.

_________________
S.B.

Steve,



I guess I missed something here.

The 1-meg resistor? What are we looking for with that?

I'm not following, what you are getting at. . .

I do know what a 100K resistor did, then went to 500K and it got worse. . .Never went to, as high as 1-meg. This was the point that I realized that a load of less than 100K was necessary.

At this time, all I know, is what I've stated in my findings.

Seeing improvements in clean audio, and audio amplitude on the scope, then hearing it, in real, on-air testing. Confirming and re-confirming what I've seen and heard is all I know at this time.

I would like to see people duplicate my experiments, then question my findings. Now, I'm telling you, that making quick tests and judgements here is not going to tell anything. So-far, I've put hours into testing these and questioning my own findings. Every time, I go back to what I keep seeing that works the best.

The cleanest, best audio that I've found is to put thses elements in series, and this is the case with resisitve, or inductive loading. Certainly, the latter is far the best, along with the more precise matching that can be had, with transformers, such as the Bogen T-725.

When I switched these elements from parallel to series, the sensitivity, and cleaner audio was very remarked. . .Not just a small amount, but very noticable.

Made the difference between hearing, or not hearing 550. This is an OPB station, low-power and in a college town, quite a few miles south of me. A difficult station to hear, from here. Everything needs to be peaked right, including the most sensitive and best headphones. On the other-side, at 1550, which is a torch, 1/2-mile from here, with everything peaked, the audio was quite clean.

The above, with parallel, and 100K resistor. . .Nothing. . .very distorted on the 1550, high-level signal. 550, was not obtainable, at all. . .This awas also the case with the elements in series and a 100K resistor. When I went to the series configuration and the 12K resistor, then it all cama alive. More life than ever, with the addition of the Bogen and the Benny, set at 33K. Which the latter, is very close to the performance of my RCA Deck-Talkers.

I hope this all helps to clarify and confirm that my findings have some merit.

Mike

_________________
I like the K-I-S-S Principle: "Keep It Sane & Simple"
http://www.peeblesoriginals.com

OK Mike, thnx!
Your findings while counter-intuitive are very informative. Thnx much for the detailed response, very helpful.
There is apparently something else going on here with these not very obvious. I'll get to follow up shortly as I found a pair of the Kyocera elements in my parts box with some others and the headset with the RS elements with shunt resistors. I'll start by taking out the shunts and playing those.

_________________
S.B.

Steve,



?????

I did drop a 1-meg resistor across the Piezo phones, and Phssst in series and doulbe Phssst in parallel. . .Just as I suspected.

Sensitivity is nowhere, and very high distortion is everywhere.

Mike

_________________
I like the K-I-S-S Principle: "Keep It Sane & Simple"
http://www.peeblesoriginals.com

Hi Mike,

Nice work.

What I don't understand is that the elements in series works better than paralleled if a transformer is used. This would mean the T-725 efficiency is better for the 5k and 10k taps than it is for the 1.2k and 2.5k taps. This could be caused by high leak inductance due to a less than ideal coupling factor. I don't have a Bogen T-725 so I can't verify that.

The 12K resistor for two elements in series makes sense. It's a high loss method to get the proper impedance match without the use of a transformer. Using a higher value resistor impairs the ability to discharge the piezo capacitance this means the amplitude decreases with increasing frequency. The piezo with parallel resistor form a low pass filter (LPF). With a 100K resistor you'll get a 47Hz LPF. With 12k this is about 390Hz. In both cases higher frequencies will be attenuated but with 12k the attenuation curve is better compensated by the piezo efficiency curve, which increases with frequency.
Lowering the resistor below 12k should improve the frequency response but due to the increased impedance mismatch, between diode and resistor, more power is wasted thus decreasing the overall sensitivity. Apparently 12k is about the optimum for two KBT-44SB-1A piezo elements in series.

Due to the large mismatch (assuming >100k detector impedance) you should be able to run two, and maybe even three or four, piezo headsets in series with only slightly weaker audio than with a single headset. This works because with two, three or four, headsets the impedance match to the detector is better and more power is available for the multiple headsets.


Corne

Mike,

You refer to the "Harbor Freight" Ear Muffs in your initial post in this thread. Are you referencing Harbor Freight p/n 43768-6VGA?

Thanks, and nice work on the piezos.

Bruce Bailey