APEM Piezo Issue

[betti154]

Hi,

I've built a dual head video solution which I'm having a few issues with and thought I'd post for comments and ideas. The solution is comprised of 5 * XMLs in series driver by a TaskLED H6Flex driver. I'm using an Apem piezo (in prolongated config) to provide on/off and low/med/hi power control.

I took it for a first dive today and am having a few issues with the the piezo. Namely the power setting jumps around a bit and it not reliable. A single push on the piezo turns the light on, then single pushes cycle between hi/med/low. A press turns the unit off. All these functions work via the piezo, but the power jumps around a bit even at constant depth. I've only used the light at around 50m, so can't say for sure if this occurs at shallower depths.

I've near zero experience with piezos and wondering if anyone had any advice. The piezos are glued into stock alloy (60mm diameter) using JB Marine Weld.

The other "feature" I noticed was that static electricity seems to trigger the piezo. This occurred in my living room on the carpet a few times, though I imagine the superlock mode would address the random turning on.

Kind Regards,
Damien Siviero

 

[350xfire]

Hi,

I've built a dual head video solution which I'm having a few issues with and thought I'd post for comments and ideas. The solution is comprised of 5 * XMLs in series driver by a TaskLED H6Flex driver. I'm using an Apem piezo (in prolongated config) to provide on/off and low/med/hi power control.

I took it for a first dive today and am having a few issues with the the piezo. Namely the power setting jumps around a bit and it not reliable. A single push on the piezo turns the light on, then single pushes cycle between hi/med/low. A press turns the unit off. All these functions work via the piezo, but the power jumps around a bit even at constant depth. I've only used the light at around 50m, so can't say for sure if this occurs at shallower depths.

I've near zero experience with piezos and wondering if anyone had any advice. The piezos are glued into stock alloy (60mm diameter) using JB Marine Weld.

The other "feature" I noticed was that static electricity seems to trigger the piezo. This occurred in my living room on the carpet a few times, though I imagine the superlock mode would address the random turning on.

Kind Regards,
Damien Siviero

Damien. That is a cool set of lights. Can you post pics of the battery set up? Can you tell us what batteries you are using. Have you taken voltage and amps measurement. How long did you run the light? Does it happen above water? Where is the driver in relation to the switch? I think I read on the Flex you have to keep the wire as short as possible to avoid interference which can switch the light through modes.

George at Taskled may be able to tell you more.

 

[Klem]

Nice lights.

I had a similar situation occur to me with the H6Flex. When connected it would switch through the modes without me pressing the piezo. Turned out to be internal damage to the driver caused by salt water. Only a few drops got past the O ring but over the course of one dive with a small pool sloshing around inside while the driver was drawing current and the damage was done.

Can also be caused by something pushing against the driver while it is sealed, like coiled cables for example.

 

[350xfire]

Here is what Taskled says:
"It is recommended to keep the maximum wire length between the switch and H6Flex less than around 8". Having wire leads that are too long can cause the wires to act as an antenna and cause H6Flex to respond irregularly or turn on/off by itself."

I think focus on what Klem and I are saying and see. Also may want to try changing the switch to a momentary toggle or something else and see.

 

[betti154]

Nice lights.

I had a similar situation occur to me with the H6Flex. When connected it would switch through the modes without me pressing the piezo. Turned out to be internal damage to the driver caused by salt water. Only a few drops got past the O ring but over the course of one dive with a small pool sloshing around inside while the driver was drawing current and the damage was done.

Can also be caused by something pushing against the driver while it is sealed, like coiled cables for example.

I'd be very surprised if water got into the housing as I test dived the heads/canister empty to 80m+ and they're dry as a bone. The first dive to 50m with electronics would suggest housing integrity too as at that pressure you don't usually get a small leak.

Can something push on the driver? For sure, space was a premium and in hindsight I should be bored out the housing more. Installing the drivers was a challenge!
There is minimal space for wiring and the piezo in the head. Piezo wires about about 4-5" and are just long enough to get the internal heatsink (which holds LED on front, driver on back) out the front.

The problem does not appear to occur in a bucket of water, though I'll run some more tests.

Come to think of it one head was reasonably stable but the other was the one bouncing around at lot, getting stuck on low power if that means anything.

 

[betti154]

Here is what Taskled says:

I think focus on what Klem and I are saying and see. Also may want to try changing the switch to a momentary toggle or something else and see.

Any suggestions for a momentary toggle design that can be used underwater. I've got a couple of Hallsw boards and was thinking about using them as a momentary switch, or even just a glass reed tube. Question I have is how easiest to make a spring loaded toggle switch - current thought is to drill a hole for some kind of sliding barrel and mill a hole into it for the sliding know. Either way, I don't think there is space in these housings to do so and I might have to leave them at fixed power and put a switch in the canister.

 

[betti154]

Damien. That is a cool set of lights. Can you post pics of the battery set up? Can you tell us what batteries you are using. Have you taken voltage and amps measurement. How long did you run the light? Does it happen above water? Where is the driver in relation to the switch? I think I read on the Flex you have to keep the wire as short as possible to avoid interference which can switch the light through modes.

George at Taskled may be able to tell you more.

Here you go.

The piezo switches are right through the back of the heads, protruding into a ~20mm cavity which has all the cables and h6flex heatsinked to the outer alloy slider (LEDs on front side of this).

Battery is 5S2P 18650s, made with 2600mah Tenergy w/ tabs (batteyrjunction.com). Pack is externally balanced and I use icharger to charge with.

Canister is made from 60mm OD alloy pipe (old kids bed tubing), wall thickness is 2mm. End is glued with Marine Weld. Top with dual glands has dual orings and is held in with two grub screws. Glands a DiveRite branded Argos's (I think), simply for availability as DiveGearExpress.com sell them and we order stuff there. Again, glands are glued in with Marine weld to remove failure points. Canister wasn't meant to be pretty, just cheap and effective.

I've run it on an watt meter, each head pulls about 2.6amp when run on a fresh charged battery (above) at full power 3000ma (configured in H6flex); LEDs fade a bit on tail end of battery but give the canister size and battery configuration I considered this acceptable (battery pack was made to made DiveRite canister size, as we use those canisters elsewhere for primaries). Wattage runs constant at about 49 watts each head.

I've got another head that runs on a 12v battery as is 3S2P LEDs, on a h6flex at 6000ma. It runs at 62watts.

oh, and balls on the heads were old Ikelite arms cut and turned down to go into heads. Again, scavenged these for free otherwise would go with a standard ball with 1/4" thread.

 

[Klem]

Again, nice settup Damien,

To be clear, I have minimal space in one of my H6Flex torches and notice the lights toggling through their programmed modes as soon as I turn it on. It only does this when I coil up the battery wiring and push it closed. It does not need to be in water to do this and I can only imagine it is either physical pressure of the battery and coiled wires pushing against the driver, or proximity of the coiled wires to the driver causing some sort of RF interference. I am thinking of building a protective disk to enclose the driver in its own compartment to stop this. Regardless, it's probably not a good thing to have wires pushing up against the driver.

I see you have 5*XM-L per head, which means the forward voltage requirement is 16.75V on max current. Plus another volt for the H6Flex and this would explain the 'tail-end' effect you are getting. I'm thinking the dimming begins when your battery reaches half-full 17.75V. I notice in my 5S2P, 4 XM-L, H6Flex torch that with high current draw there is also a voltage drop in the battery, but then again I am using cheapy Trustfire's. The H6Flex won't take another cell in that battery however so unfortunately it's not a quick fix.

One thought, if you want those lights on max brightness for the full life of the battery you could go 4*XM-L U2, instead of 5*XM-L T6. The newer U2 can take 3.5A and delivers 1,200lumen. 5K lumens for half the battery with dimming as opposed to 4.8K for the full battery.

Apologies if you know this already.

 

[350xfire]

Very cool indeed. 49 watts per head. That's monstrous! I bet out of the water you get about 40 seconds before it boils. I just built a 3XML and they do get hot!
I was wondering why the raised spot on the back of the head. Makes sense now. So distance is not an issue!
I sent George an email to check this out.

 

[betti154]

I was wondering why the raised spot on the back of the head. Makes sense now. So distance is not an issue!
I sent George an email to check this out.

The raised spot on the back of the heads is the Apem piezo, and more a product of the Apem profile than anything else. We originally made the housing to fit the Shurter piezo which are 6-7mm shorter in length I think. Space certainly became a premium somewhere along the way.

Here is a very crude and not to scape diagram of the assembly. You should get the idea, particularly how it relates to the h6flex and space.

Based on what I've heard I've some relief in that people seem to think it's my installation of the h6flex and not the Apem piezos that are the issue. My initial reaction after the dive was incompatible piezo, which is annoying as I've glued them in and have another 4 (had to bulk order).

I'll wait until I heard form George, but currently thinking about boring more of the heatsink out to make more room. Of course this is a balance as I'm eating into the heatsink, but no matter I've choice I guess.

Is there anything I can use for insulation?

 

[georges80]

Wire length from the switch to the driver can be a problem, especially if it picks up EMI/RFI from nearby sources - even self inflicted through the power input leads or LED output leads... Having power leads and LED leads bunched up close against the switch wiring is NOT a good idea, since those power/LED leads can carry quite a bit of current that have be 'noisy' on some of the lower settings where the h6flex uses PWM for dimming (below 1A).

In the OP case you need to determine if the piezo is sending a true 'on' signal when at depth in water - i.e. the piezo is sensing the external water pressure as a press or whether it is an EMI/RFI pickup issue.

If this only occurs at depth, then it seems reasonably likely that the piezo is actually sending the 'on' to the h6flex and it is responding to that correctly.

The input circuitry of the switch input is basically a pullup resistor of 10k ohms (in parallel with a weak internal pullup inside the microcontroller). This is pulled up to nominal 3V (for the h6flex). You could place a capacitor in parallel with the SWA/SWB inputs of say 0.01uF (even up to 0.1uF) to create an RC filter. This would prevent the switch input from seeing short spikes/pulses that may be coupling onto the wire.

cheers,
george.

 

[betti154]

Again, nice settup Damien,

To be clear, I have minimal space in one of my H6Flex torches and notice the lights toggling through their programmed modes as soon as I turn it on. It only does this when I coil up the battery wiring and push it closed. It does not need to be in water to do this and I can only imagine it is either physical pressure of the battery and coiled wires pushing against the driver, or proximity of the coiled wires to the driver causing some sort of RF interference. I am thinking of building a protective disk to enclose the driver in its own compartment to stop this. Regardless, it's probably not a good thing to have wires pushing up against the driver.

I see you have 5*XM-L per head, which means the forward voltage requirement is 16.75V on max current. Plus another volt for the H6Flex and this would explain the 'tail-end' effect you are getting. I'm thinking the dimming begins when your battery reaches half-full 17.75V. I notice in my 5S2P, 4 XM-L, H6Flex torch that with high current draw there is also a voltage drop in the battery, but then again I am using cheapy Trustfire's. The H6Flex won't take another cell in that battery however so unfortunately it's not a quick fix.

One thought, if you want those lights on max brightness for the full life of the battery you could go 4*XM-L U2, instead of 5*XM-L T6. The newer U2 can take 3.5A and delivers 1,200lumen. 5K lumens for half the battery with dimming as opposed to 4.8K for the full battery.

Apologies if you know this already.

Thanks for the info - to be honest I find the LED binning confusing as hell. I'm using these LEDs from cutter to be specific as I liked the 4500K colour for the green-ish tinge water we come across in Eastern NSW - http://www.cutter.com.au/search.php?pg=1&stext=Cutter-XML14RNDE5&sprice=&stype=&scat=&sman=

 

[betti154]

Wire length from the switch to the driver can be a problem, especially if it picks up EMI/RFI from nearby sources - even self inflicted through the power input leads or LED output leads... Having power leads and LED leads bunched up close against the switch wiring is NOT a good idea, since those power/LED leads can carry quite a bit of current that have be 'noisy' on some of the lower settings where the h6flex uses PWM for dimming (below 1A).

In the OP case you need to determine if the piezo is sending a true 'on' signal when at depth in water - i.e. the piezo is sensing the external water pressure as a press or whether it is an EMI/RFI pickup issue.

If this only occurs at depth, then it seems reasonably likely that the piezo is actually sending the 'on' to the h6flex and it is responding to that correctly.

The input circuitry of the switch input is basically a pullup resistor of 10k ohms (in parallel with a weak internal pullup inside the microcontroller). This is pulled up to nominal 3V (for the h6flex). You could place a capacitor in parallel with the SWA/SWB inputs of say 0.01uF (even up to 0.1uF) to create an RC filter. This would prevent the switch input from seeing short spikes/pulses that may be coupling onto the wire.

cheers,
george.

Thanks for the input George!

What I find most interesting is that one head is significantly problematic and the other more almost perfect. This suggests the piezo may be ok, and it is my crude wiring and assembly that are causing the problem. I'll run more bucket tests and take for another dive before determining the issue.

I think I'll even take the housing for a dive with the piezo wired to a small battery and status LED. I should simply be able to watch the LED underwater and see if the apem piezo is closing the circuit in response to depth change or other events.

 

[Flo1]

Hey Damien,

That's a PITA!

I'll have to see if I have any issues with the maxflex setup I use..

Cheers

Flo

 

[Flo1]

Wire length from the switch to the driver can be a problem, especially if it picks up EMI/RFI from nearby sources - even self inflicted through the power input leads or LED output leads... Having power leads and LED leads bunched up close against the switch wiring is NOT a good idea, since those power/LED leads can carry quite a bit of current that have be 'noisy' on some of the lower settings where the h6flex uses PWM for dimming (below 1A).

In the OP case you need to determine if the piezo is sending a true 'on' signal when at depth in water - i.e. the piezo is sensing the external water pressure as a press or whether it is an EMI/RFI pickup issue.

If this only occurs at depth, then it seems reasonably likely that the piezo is actually sending the 'on' to the h6flex and it is responding to that correctly.

The input circuitry of the switch input is basically a pullup resistor of 10k ohms (in parallel with a weak internal pullup inside the microcontroller). This is pulled up to nominal 3V (for the h6flex). You could place a capacitor in parallel with the SWA/SWB inputs of say 0.01uF (even up to 0.1uF) to create an RC filter. This would prevent the switch input from seeing short spikes/pulses that may be coupling onto the wire.

cheers,
george.

Just wondering whether shielding the wire pairs would help..?

 

[betti154]

Hi All,

My approach to finding a solution to this problem is two part, and involves the housing heads that I've already built.

1. In Head 1, I'm going to install a 0.1uf capacitor between SWA/SWB and see if that cleans up the signal.
2. In head 2, I'll install a simple status LED circuit and wire it in series with the piezo in the head. The idea is that the piezo closes the circuit (powered by a small 3.7v battery in head) and turns on the 5mm red LED, allowing me to visually test piezo function (both click and press) independent of the h6flex at depth. I should also be able to observe whether the piezo is closing in response to depth changes.

Thoughts, comments and advice welcome?

My question with the capacitor is whether this is as simple as getting a 100NF capacitor and soldering it into SWA/SWB pins? I think take the piezo leads and wire them to either size of the capacitor? Is polarity an issue? Excuse my primitive understanding of electronics as I'm a photographer/computer guy and this is stretching my knowledge.

thanks, damien

 

[Flo1]

Hi Damien,

I reckon before you go and put the capacitor in you should give the shielding idea a go. By shielding either (or both) the switching leads or power leads you should be able to filter any stray EMF/RFI and perhaps solve your issue. Just make sure you ground the shielding to make it most effective.

Cheers.

Flo

 

[350xfire]

Damien:
Are you running a driver per head? If so, are you running off the same battery?

George, will your drivers play well in series or parallel?

 

[betti154]

Hi Damien,

I reckon before you go and put the capacitor in you should give the shielding idea a go. By shielding either (or both) the switching leads or power leads you should be able to filter any stray EMF/RFI and perhaps solve your issue. Just make sure you ground the shielding to make it most effective.

Cheers.

Flo

Shield with what?

 

[betti154]

Damien:
Are you running a driver per head? If so, are you running off the same battery?

George, will your drivers play well in series or parallel?

There is a single battery with a driver per head running at 3000ma. Drivers are wired in parallel off battery (no on/off switch other than h6flex). drivers are in ubi2 mode.

One interesting point is that on power up (where the drivers flash for a few seconds), one head flashes a few seconds longer before going steady. I'm not sure how to explain this since they're configured identically.