no, everything needs to be powered by 1 battery pack.
Can someone recommend a level shifter to purchase off digikey and a schematic please?
I did some searching and it seems the TC4427 is popular for this purpose.
I searched digikey but got back almost 100 different results, all mosfet's
I've been using a now discontinued level shifter from SparkFun. I think this is their new version.
JonnyMac uses the TC4427 but IIRC they're over $1 each for low quantities. I think there are likely less expensive options. I'm hoping some sort of simple (and cheap) transistor would work but I haven't tried this yet.
I think the chip used to control the LED is WS2811 while the LEDs with these chips inside called WS2812 or WS2812B for the four lead version.
Edit: If I were to purchase a TC4427 from Digi-Key, I'd probably go with this one. It's just barely over a dollar at $1.01.
ill probably go ahead and get one of those pre-assembled from sparkfun... but if that proves to be the solution, i would like to make my own using the TC4427.
Do you have a schematic by chance of how to wire it up for this purpose?
I looked through the datasheet but theres nothing in there.... sorry, im completely new to having to use a level shift
With my 72 LED strip I had no issues running 3.3v logic. I guess each LED sinks a little more current on the IO line, and eventually it's too much.
If that is the case, I still don't see how that could have to do with a change in the power source. I would think it has to do more with noise on your vin and vdd.
Are you sure the regulators you're using with the assortment of batteries are not the source of the problem?
Can you drive the battery regulator off your bench power supply instead of a battery and test the setup?
Also I keep ignoring the battery compartment allocation. Are you really set on this size? I think your choice of lipo batteries was good, but you're going to have to program your routines to limit current. The best way to do that is to use a high current source that you can monitor closely for the initial programming. A few strips of white running off a battery with the dimensions you listed will not last long. You will need a good under volt cutoff buffer because the load will be going from a few hundred ma to amps during the routine. The resting voltage may not be a good indicator of the pack health.
I would consider running a 11.1v 3S 2000mah 40c (continuous) battery and a 5 amp DC to DC converter with a low voltage cutoff. This should get you around 20 minutes of eye watering white, and leave you some wiggle room on current.
Lawson:
I sent a message to the guys at cheapbatterypacks.com to see if they can offer a solution.
I did not know about the difference between high mah rating and high discharge burst... I thought they were the same thing.
I wish I had a way to compare the eneloop 950mah batteries with those KAN 2/3AAA's... I also tried energizer 900mah AAA's with the same results.
Thanks for the info, hopefully they can provide a solution.
They also have a KAN AAA cell if you want an exact drop in for your current cells. The 10C rating is almost always for discharge, and different manufacturers use different testing criteria so It's hard to compare voltage droop under load between manufacturers without seeing detailed performance plots. (or just testing) Though the 2/3AAA size should be short enough to let you use 5 cells and a small 5V switching regulator if the LED strings don't work well at less than 5v. I.e. While NiMh cells will be 1.5 to 1.6 volts each right off the charger, they'll quickly drop to about 1.2 volts under load and are "empty" when they reach 0.9 volts or so.
The guys at cheapbatterypacks were very helpful and recommended 5x KAN700 AAA NiMH batteries... they were able to add nickle tabs to them for easy soldering.
At $3/ea thats a deal if they work as good as they say. I paid something like $25 for 4 of the eneloop batteries.
I also ordered the level converter from sparkfun. I see its really just a simple TTL type device. I have made similar circuits on laser power drivers with TTL inputs by using both transistors and opto-isolators.
Thanks for pointing out the level converter... i did not think about the voltage difference at all
Also on the way are MCP1700's and LP2950's to replace the LM2937 3.3v regulator.
This is looking real promising for resolving the issue. I will post results back asap.
Thanks so much for the help guys!
Received the Digikey parts today. Still waiting on the level converter and KAN batteries.
Tested out both the MCP1700 and the LP2950.
the LP2950 came with an odd warning about humidity that I have never seen before, even had a humidity indicator card inside the package with a note to bake @ 125deg if indicator showed above 60%rh. Not sure if this means it will malfunction if it ever is used in high humidity? or if this was just a storage concern? I could find no such warnings in the datasheet.
Both the MCP and LP seemed to resolve the problem (knock on wood).
I briefly tested 200 LED's on both regulators for about 10 minutes without trouble.
I then made sure the batteries had a full charge (still using the 4x eneloop 950mah nimh AAA's), and ran 200 LED's using the LP2950 for one hour and 15 minutes straight before the batteries needed recharging.
I am now recharging the batteries again and will do another endurance run later today using the MCP1700 to see if there is a difference in run-time.
The MCP1700 outputs 250ma, the LP2950 outputs 100ma... so im expecting the LP to last longer, my only concern is the humidity warning on the packaging.
Has anyone else seen this warning before? The datasheet for the LP part says it is military grade, so I dont know if it's just a requirement or if it will malfunction in high humidity, which it will definitely be exposed to during the summer.
I
So I guess the main problem was the LM2937 sucking 500ma from the batteries? That is about 60% of the rated mah of the batteries.
I still plan to do testing with both the level converter and the KAN batteries when they arrive.
I personally never worry about them but then I don't make more than a few boards at a time. I think the humidity warning applies to most electronic components but Digi-Key just spins a roulette wheel to decide which packages to include the cards.
I've had parts outside of their special humidity sealed packages for several years and I haven't had trouble when I finally get around to adding them to a PCB. I do live in a dry area though.
As far as I know those humidity warnings are assuming you're going to re-flow solder the parts. I.e. quickly heat the whole part up to over 183C. I would assume that at re-flow soldering temperatures and heating rates, excess water absorbed in the package can break the package.
Since the LP2950 fixed the problem, I say that most likely the LM2937 was oscillating and getting the prop to hit the 2.7v brownout voltage too soon. I'd expect the MCP1700 to give the same run-time as the LP2950 unless is has a lower dropout voltage. The current ratings for the MCP1700 and LP2950 are just nominal limits, your circuit determines the actual current draw.
Ah thats makes good sense about the humidity and high heat combo. Ive ordered thousands of IC's from digikey but had never seen this before, wasnt sure if it was a unique issue with this part.
The LP2950 does run longer than the MCP1700.
I ran two 'lifetime' tests on each of the two regulators... so four tests total.
After recharging 4x AAA's, the LP2950 ran for 1h 15min, the MCP1700 ran 55min. These results were consistent in both tests on each regulator.
It seems to me that even though the prop isnt using the max supply current of either regulator... each regulator pulls from the batteries its max supply current... this is the only way I can account for the difference in running time.
I may be wrong, but why else would there be a 20min difference in run time off the same batteries? The LP output = 100ma, the MCP output = 250ma
The level converter arrived today.
I am testing it right now using the LP2950 regulator and will compare the lifetime again.
Been running now for 18min and looks good... no difference however than without it.
Just curious what the difference in running time will be with the level converter in use.
Hopefully I will not need to use it.
Ah thats makes good sense about the humidity and high heat combo. Ive ordered thousands of IC's from digikey but had never seen this before, wasnt sure if it was a unique issue with this part.
The LP2950 does run longer than the MCP1700.
I ran two 'lifetime' tests on each of the two regulators... so four tests total.
After recharging 4x AAA's, the LP2950 ran for 1h 15min, the MCP1700 ran 55min. These results were consistent in both tests on each regulator.
It seems to me that even though the prop isnt using the max supply current of either regulator... each regulator pulls from the batteries its max supply current... this is the only way I can account for the difference in running time.
I may be wrong, but why else would there be a 20min difference in run time off the same batteries? The LP output = 100ma, the MCP output = 250ma
The level converter arrived today.
I am testing it right now using the LP2950 regulator and will compare the lifetime again.
Been running now for 18min and looks good... no difference however than without it.
Just curious what the difference in running time will be with the level converter in use.
Hopefully I will not need to use it.
Huh, that run time difference is odd. The MCP1700's input can get about 0.2 volts closer to the output voltage before loosing regulation, so I'd expect it to run several minutes longer. Do you have a 1uF or larger ceramic capacitor right at the output of the regulator? (I assume you do since the LP2950 needs >3.3uF to be stable at rated current)
Hm... the MCP1700 uses a P-FET at it's output, while the LP2950 uses a PNP transistor. FETs conduct well in both directions, while PNP transistors hardly pass any reverse current. I think the LP2950 lasts longer because when the battery voltage dips below 3.3v due to LED switching, the LP2950 acts like a (poor) diode and lets the Prop's supply ride through the few micro-second interruption. Btw. explicit reverse current blocking can be found on several regulators. Off hand the MIC5232 is the only one I know of though. (the 10mA rating isn't enough for a prop at full speed)
yes I have a 10uf cap directly on the output pins of both the LP and MCP as recommended in the datasheet... there are two .1uf caps later in the circuit close to the voltage inputs of the prop.
Got a reduced runtime of 10 min with the level converter.
I did notice something that I had not payed much attention to previously.... before using the level converter, there was a very tiny 'quivering' in the led display, like as if each led was flickering but just barely enough to notice only if you looked very closely.
With the level converter, that quivering is gone and the output is very smooth.
Duane:
With the requirement of the prop be running at 80mhz, will it still work fine using the 10mhz crystal but still operating at 80 mhz by doing this:
_clkmode = xtal1 + pll8x
_xinfreq = 10_000_000
There is a note that was in the code as follows:
_clkmode = xtal1 + pll16x ' 16x required for WS2812
Also, im curious what IC chip is being used as the WS2811 controller?
I have searched digikey for WS2811 but get no results.
I would like to build some of my own strips with 5mm RGB 4-pin led's, but cant figure out what the IC is for WS2811 control
I'll have to look around for them... that is the right chip, but 50c is a bit much considering I need quantities in the hundreds. I could get a 100led strip with the ws2811 chip included for less than it would cost to buy just the ws2811 chips from that site.
Maybe I can find them on alibaba or somethin
I believe I am going to mark this thread as solved!
Thanks so much for the help guys... I would have never figured this out without the incredible knowledge of small details like the LM2937 oscillating, which I believe was indeed the cause of the problem. Both the MCP1700 and LP2950 work like a charm.
By far the greatest feature of using the propeller is the forum
Thank you all!
Comments
Can someone recommend a level shifter to purchase off digikey and a schematic please?
I did some searching and it seems the TC4427 is popular for this purpose.
I searched digikey but got back almost 100 different results, all mosfet's
JonnyMac uses the TC4427 but IIRC they're over $1 each for low quantities. I think there are likely less expensive options. I'm hoping some sort of simple (and cheap) transistor would work but I haven't tried this yet.
I think the chip used to control the LED is WS2811 while the LEDs with these chips inside called WS2812 or WS2812B for the four lead version.
Edit: If I were to purchase a TC4427 from Digi-Key, I'd probably go with this one. It's just barely over a dollar at $1.01.
Do you have a schematic by chance of how to wire it up for this purpose?
I looked through the datasheet but theres nothing in there.... sorry, im completely new to having to use a level shift
They're very easy to hook up because of the pin layout and each pin is marked. I used this guide for that board - https://learn.sparkfun.com/tutorials/using-the-logic-level-converter
With my 72 LED strip I had no issues running 3.3v logic. I guess each LED sinks a little more current on the IO line, and eventually it's too much.
If that is the case, I still don't see how that could have to do with a change in the power source. I would think it has to do more with noise on your vin and vdd.
Are you sure the regulators you're using with the assortment of batteries are not the source of the problem?
Can you drive the battery regulator off your bench power supply instead of a battery and test the setup?
I would consider running a 11.1v 3S 2000mah 40c (continuous) battery and a 5 amp DC to DC converter with a low voltage cutoff. This should get you around 20 minutes of eye watering white, and leave you some wiggle room on current.
They also have a KAN AAA cell if you want an exact drop in for your current cells. The 10C rating is almost always for discharge, and different manufacturers use different testing criteria so It's hard to compare voltage droop under load between manufacturers without seeing detailed performance plots. (or just testing) Though the 2/3AAA size should be short enough to let you use 5 cells and a small 5V switching regulator if the LED strings don't work well at less than 5v. I.e. While NiMh cells will be 1.5 to 1.6 volts each right off the charger, they'll quickly drop to about 1.2 volts under load and are "empty" when they reach 0.9 volts or so.
Marty
At $3/ea thats a deal if they work as good as they say. I paid something like $25 for 4 of the eneloop batteries.
I also ordered the level converter from sparkfun. I see its really just a simple TTL type device. I have made similar circuits on laser power drivers with TTL inputs by using both transistors and opto-isolators.
Thanks for pointing out the level converter... i did not think about the voltage difference at all
Also on the way are MCP1700's and LP2950's to replace the LM2937 3.3v regulator.
This is looking real promising for resolving the issue. I will post results back asap.
Thanks so much for the help guys!
Tested out both the MCP1700 and the LP2950.
the LP2950 came with an odd warning about humidity that I have never seen before, even had a humidity indicator card inside the package with a note to bake @ 125deg if indicator showed above 60%rh. Not sure if this means it will malfunction if it ever is used in high humidity? or if this was just a storage concern? I could find no such warnings in the datasheet.
Both the MCP and LP seemed to resolve the problem (knock on wood).
I briefly tested 200 LED's on both regulators for about 10 minutes without trouble.
I then made sure the batteries had a full charge (still using the 4x eneloop 950mah nimh AAA's), and ran 200 LED's using the LP2950 for one hour and 15 minutes straight before the batteries needed recharging.
I am now recharging the batteries again and will do another endurance run later today using the MCP1700 to see if there is a difference in run-time.
The MCP1700 outputs 250ma, the LP2950 outputs 100ma... so im expecting the LP to last longer, my only concern is the humidity warning on the packaging.
Has anyone else seen this warning before? The datasheet for the LP part says it is military grade, so I dont know if it's just a requirement or if it will malfunction in high humidity, which it will definitely be exposed to during the summer.
I
So I guess the main problem was the LM2937 sucking 500ma from the batteries? That is about 60% of the rated mah of the batteries.
I still plan to do testing with both the level converter and the KAN batteries when they arrive.
I personally never worry about them but then I don't make more than a few boards at a time. I think the humidity warning applies to most electronic components but Digi-Key just spins a roulette wheel to decide which packages to include the cards.
I've had parts outside of their special humidity sealed packages for several years and I haven't had trouble when I finally get around to adding them to a PCB. I do live in a dry area though.
Since the LP2950 fixed the problem, I say that most likely the LM2937 was oscillating and getting the prop to hit the 2.7v brownout voltage too soon. I'd expect the MCP1700 to give the same run-time as the LP2950 unless is has a lower dropout voltage. The current ratings for the MCP1700 and LP2950 are just nominal limits, your circuit determines the actual current draw.
Marty
The LP2950 does run longer than the MCP1700.
I ran two 'lifetime' tests on each of the two regulators... so four tests total.
After recharging 4x AAA's, the LP2950 ran for 1h 15min, the MCP1700 ran 55min. These results were consistent in both tests on each regulator.
It seems to me that even though the prop isnt using the max supply current of either regulator... each regulator pulls from the batteries its max supply current... this is the only way I can account for the difference in running time.
I may be wrong, but why else would there be a 20min difference in run time off the same batteries? The LP output = 100ma, the MCP output = 250ma
The level converter arrived today.
I am testing it right now using the LP2950 regulator and will compare the lifetime again.
Been running now for 18min and looks good... no difference however than without it.
Just curious what the difference in running time will be with the level converter in use.
Hopefully I will not need to use it.
Huh, that run time difference is odd. The MCP1700's input can get about 0.2 volts closer to the output voltage before loosing regulation, so I'd expect it to run several minutes longer. Do you have a 1uF or larger ceramic capacitor right at the output of the regulator? (I assume you do since the LP2950 needs >3.3uF to be stable at rated current)
Hm... the MCP1700 uses a P-FET at it's output, while the LP2950 uses a PNP transistor. FETs conduct well in both directions, while PNP transistors hardly pass any reverse current. I think the LP2950 lasts longer because when the battery voltage dips below 3.3v due to LED switching, the LP2950 acts like a (poor) diode and lets the Prop's supply ride through the few micro-second interruption. Btw. explicit reverse current blocking can be found on several regulators. Off hand the MIC5232 is the only one I know of though. (the 10mA rating isn't enough for a prop at full speed)
Marty
Got a reduced runtime of 10 min with the level converter.
I did notice something that I had not payed much attention to previously.... before using the level converter, there was a very tiny 'quivering' in the led display, like as if each led was flickering but just barely enough to notice only if you looked very closely.
With the level converter, that quivering is gone and the output is very smooth.
With the requirement of the prop be running at 80mhz, will it still work fine using the 10mhz crystal but still operating at 80 mhz by doing this:
There is a note that was in the code as follows:
Also, im curious what IC chip is being used as the WS2811 controller?
I have searched digikey for WS2811 but get no results.
I would like to build some of my own strips with 5mm RGB 4-pin led's, but cant figure out what the IC is for WS2811 control
WS2811 is the name of the controller.
Here are some PCBs with just the WS2811 chip without LEDs. I'm pretty sure it's possible to purchase the chips on their own.
that is the same thing i was looking at, thought i might be able to find a p/n on one of the pics.... http://www.ebay.com/itm/10-pcs-x-9mm-WS2811-Circuit-Board-PCB-for-making-5V-12mm-WS2811-Led-pixel-module/251358402644?
but it says:
WORLDSEMI
WS2811F
20120201
I searched digikey for any of those including just '2811' but didnt get any results.
Found just the chip here: http://www.ebay.com/itm/100PCS-WS2811S-WS2811-SOP-8-WORLDSEMI-CHIP-IC-/141105187653?_trksid=p2054897.l4276
And found a datasheet here: http://www.adafruit.com/datasheets/WS2811.pdf
This datasheet doesnt show the difference between the WS2811F or WS2811S.... and shows an input voltage of 6-7v and an output voltage of 12 which is weird, why would it say 12v out?
http://www.adafruit.com/products/1378?gclid=CIfYl7PUh70CFYN0Ogod4g0ABg
Maybe I can find them on alibaba or somethin
Thanks so much for the help guys... I would have never figured this out without the incredible knowledge of small details like the LM2937 oscillating, which I believe was indeed the cause of the problem. Both the MCP1700 and LP2950 work like a charm.
By far the greatest feature of using the propeller is the forum
Thank you all!