« first day (2275 days earlier)      last day (2658 days later) » 

Giskard42
Giskard42
12:11 AM
@Skyler
I have a rig now working with two 640*480 cams
 
Giskard42
Giskard42
12:24 AM
@Skyler Seems to work okay. ~33 percent CPU usage, RPI3, unoverclocked, 15 fps. Some frames are dropped, I believe because I'm storing to a slow SD card.
Doesn't even hit the swap
 
Giskard42
Giskard42
12:34 AM
@Skyler Haven't tried displaying those videos, trying that now.
Oh, and that's 33 percent CPU per cam btw
 
Eric Urban
Eric Urban
@Giskard42 i guess i need to order some of those specialized scope probes. I've definitely found that the orientation of the ground clip can change the measured signal
Can a MAX5075 replace a TL494? I'm trying to get above the 90% duty cycle limit of the TL494 in a push pull driver
 
Tony Stewart
Tony Stewart
12:53 AM
not often one gets to save a "life" here electronics.stackexchange.com/questions/281126
@EricUrban hve you tried removing probe tip and gnd wire and just use tip and barrel between to test pins (sig,gnd)
 
Eric Urban
Eric Urban
tip and barrel? I've not done that. I don't know how to take the probes apart
 
Tony Stewart
Tony Stewart
usually just pulls off
that is the best probe method
5mm pin and shield barrel
 
Eric Urban
Eric Urban
ill have to try that time next time I get to play with the circuit. My other idea was to rearrange things onto some sheet metal to minimize the circuit length between the inductor, capacitor, and board
 
Tony Stewart
Tony Stewart
use twist pairs for signals fast signal edges using gnd and magnet wire
or utp wire
 
Eric Urban
Eric Urban
i guess I should photograph the copper side of the board
I made the length for the signals that control the MOSFETs super short
 
Tony Stewart
Tony Stewart
12:58 AM
keep inductive current loop areas SMALL
 
Eric Urban
Eric Urban
i did pretty well in that regard I think. no more gate bounce at least.
 
Tony Stewart
Tony Stewart
with probe tip and ground removed, I usually add test point pairs or get the mini spring probe for gnd.
 
Eric Urban
Eric Urban
i will have to see if the CAD software I used can do that
 
Tony Stewart
Tony Stewart
advanced methods use a FET buffered diff probe
 
Eric Urban
Eric Urban
the biggest annoyance at present is realizing my TL494 only can do 90% duty cycle
so if a MOSFET breaks down gate-to-source it is usually destroyed instantly, but drain-to-source it seems to survive?
 
Tony Stewart
Tony Stewart
1:03 AM
" The output pulse width varies from 97% of the period to 0 as the voltage present at the error amplifier output varies from 0.5 V to 3.5 V, respectively. " So why is yours limited to 90%?
 
Eric Urban
Eric Urban
correct, thats the oscillator. The output transistors have a 'typical' duty cycle of 90% according to the data sheet however.
 
Tony Stewart
Tony Stewart
it says that for PWM
 
Eric Urban
Eric Urban
on the datasheet, in section 7.9
 
Tony Stewart
Tony Stewart
This requires the control signal (error amplifier output) to be ∼0.7 V greater than the voltage across CT to inhibit the output logic, and ensures maximum duty cycle operation without requiring the control voltage to sink to a true ground potential.
 
Eric Urban
Eric Urban
2nd row onf the table "Maximum duty cycle" is 45%
 
Tony Stewart
Tony Stewart
1:08 AM
ah 10% min deadtime
 
Eric Urban
Eric Urban
that was my conclusion. At very low freq. for some reason I've seen it do a 100% duty cycle
but that was @ 130 Hz
not very common freq. for SMPS
 
Tony Stewart
Tony Stewart
"An internal offset of 110 mV ensures a minimum dead time of ∼ 3% with the dead-time control input grounded.
 
Eric Urban
Eric Urban
yeah i cant make sense of the either. The datasheet seems to contradict itself
i didnt pay anything for the TL494s, so not really a big deal
 
Tony Stewart
Tony Stewart
1:27 AM
Considering the chip is still around after 30+ yrs, can't be too much wrong with it.
 
Eric Urban
Eric Urban
i see it used in single-ended/flyback configuration power supplies alot
where dead-time/duty-cycle is not really an issue if I understand correcltly
mcc33025 has totem pole outputs and is $3 each from digikey
 
 
2 hours later…
Skyler
Skyler
3:48 AM
@Giskard42 sorry was in lab, did you have a third webcam?
I'm curious to see if you can push it to that
 
 
3 hours later…
Krishn Shweta
Krishn Shweta
6:57 AM
OMG!!! I've never seen @Asmyldof is off from this chat room.
 
abdullah kahraman
abdullah kahraman
lol
 
Krishn Shweta
Krishn Shweta
@abdullahkahraman Is there any reason he's off?
 
abdullah kahraman
abdullah kahraman
dunno
 
Krishn Shweta
Krishn Shweta
hmm... If he won't come back then certainly this room will be silent most of the times (IMO)
 
abdullah kahraman
abdullah kahraman
He was here yesterday
 
Krishn Shweta
Krishn Shweta
7:13 AM
Yeah his last seen is 1 day ago.
 
 
2 hours later…
PlasmaHH
PlasmaHH
8:45 AM
He is playing with a new shipment of stuff I guess
 
PlasmaHH
PlasmaHH
9:10 AM
400€ shipping costs from the US for some counter... wtf are they using for shipment? a taxi?
 
 
7 hours later…
Extrarius
Extrarius
4:04 PM
I'd like to implement "Robust, low-cost, auditable random number generation for embedded system security" using through-hole components, but I can't find a good through-hole replacement for the voltage booster TPS61041. Any suggestions?
 
Giskard42
Giskard42
4:21 PM
@Skyler I do have a third webcam, I'm going to try with that today.
I don't really think there's any issue
Even with re-encoding and then saving the data to an SD card, it works fine CPU-wise
So I don't think that's a problem
I wasn't able to get a gui working on my pi, so I haven't tried visualizing and rendering live
I'll try that today
 
AndyD273
AndyD273
4:47 PM
I am curious... I am trying to figure out a way to control temperature on a TEC1-12706 peltier to keep the hot side from getting too hot. I've seen a suggestion for using a 555 timer to control temperature, but would it be possible to use a thermrister and a transistor to do the same thing? Basically have it turn off when the hot side gets to a specific temperature?
Also, are there any problems with under driving a peltier? the TEC1-12706 is supposed to be 12V, but I've run it off of a 9V battery and the hot side gets pretty hot. I've thought about just using a resistor to keep it from getting that hot, but apparently it doesn't work that way...
Sorry if these are stupid questions.
 
Giskard42
Giskard42
5:40 PM
@AndyD273 You may find this interesting: electronics.stackexchange.com/questions/28634/…
@AndyD273 I think your direct thermistor-transistor idea is a good one, actually.
The transistor is going to be dissipating a lot of power, but a big heatsink should make that managable.
I don't know if this'll work out, but you could try putting a thermistor to positive, a potentiometer to ground to adjust set temperature
You'll need a beefy transistor in to-220 or so
NPN would be best
The beauty of your analog thermistor-transistor design is that you completely ignore the difficulties of filtering pwm from an arduino
 
AndyD273
AndyD273
@Giskard42 What I'm hoping to do with it is generate heat, but not that much. Maybe limit it to 100F-105F. Basically something like this, but for heating instead of cooling
 
Giskard42
Giskard42
@AndyD273 Are you sure you need a peltier for that?
Why not a big sand resistor?
 
AndyD273
AndyD273
Maybe not, but then it could be reversible by switching the polarity
 
Giskard42
Giskard42
They're much more tolerant of PWM
@AndyD273 Hmm, okay
 
AndyD273
AndyD273
Seems like I could get a DPDT switch and it could be wired for heating or cooling...
 
Giskard42
Giskard42
5:50 PM
@AndyD273 Sure
 
AndyD273
AndyD273
I've got some 2n2222 and 2n3904 transistors, so I'm hoping one of them would work.
But if a thermister loses resistance when it gets warm, wouldn't that make the npn transistor more saturated as it warmed up?
Or do I have that backward again?
 
Giskard42
Giskard42
@AndyD273 Okay, so you put the thermistor to ground and the pot to positive
Good point
Yeeaaah, that's not going to work
witha 2n222
or a 2904
*3904
Worst case scenario, you'll be dissipating a few watts
Those are going to catch fire, likely
@AndyD273 Those also have a current limit of 0.2 amps, too
And 0.8 amps for the 2222
 
AndyD273
AndyD273
Right, but a 9V only puts out 500mA (so like .5A?)
(Sorry I really am new to this stuff)
 
Giskard42
Giskard42
@AndyD273
whoops
 
AndyD273
AndyD273
np
 
Giskard42
Giskard42
6:01 PM
@AndyD273 A 9v is good for very little current, but the peltier says it draws 100watts
You might get a little bit of heat out of it, but I have doubts you'll get any cooling at all from a 9v.
But I may be wrong.\
 
AndyD273
AndyD273
All I know is that it gets too hot when I put a 9V on it. I don't know how long the battery will last like that.
 
Giskard42
Giskard42
@AndyD273 Huh. Guess I'm wrong.
@AndyD273 Okay, so that changes things up a bit. Your 2n3904s might work fine then
 
AndyD273
AndyD273
It might not be practical, or I may have to use a differnt kind of battery pack. This is a kind of proof of concept
OK, I can at least try it, if the 9V won't make the transistor explode
 
Giskard42
Giskard42
@AndyD273 Not immediately, but it will get pretty hot
 
AndyD273
AndyD273
Probably use the 2222 since it has higher mA handling
I kinda wish Digikey was a place I could go visit. instead of having to pay shipping :/
 
Giskard42
Giskard42
6:06 PM
@AndyD273 That peltier says it has a 2 ohm resistance, so it's going to be drawing about 3 amps from that 9v battery, ideally
@AndyD273 9v batteries have a capacity of 500mah, so you'll get like 10 minutes of useful life out of it or so
I'm neglecting internal resistance, you're not going to realistically get 3a out of one of those batteries
But hey, science is about experiment. Can't hurt to try!
 
AndyD273
AndyD273
Down the road I might see about wiring it to a different battery pack. Maybe an old cell phone battery or something.
Then it could be rechargable
@Giskard42 So long as I don't burn the house down
 
Giskard42
Giskard42
@AndyD273 I'd be a bit wary about subjecting lithium batteries to that much current, even for a short time
You may get that "house burnt down" effect.
 
AndyD273
AndyD273
True. I really have a lot of reading to do...
@Giskard42 If I end up running it with a thermister and transistor to regulate power, do I need to worry about PWM, since there is no micro controller involved?
 
Giskard42
Giskard42
@AndyD273 Nope, that's the beauty of using it that way
It's an entirely linear supply
 
AndyD273
AndyD273
Cool, which is supposed to be more efficient for the peltier (from that link you posted)
Say I want to try to simulate the circuit before actually putting it together... Is there a preferred web site for that?
 
Giskard42
Giskard42
6:19 PM
@AndyD273 I think that would be difficult.
There are electronics simulators, (LTspice et al)
And there are thermal simulators
But combining them to see how the feedback works would be difficult
I would recommend using the greatest simulator ever made,
the universe
:P
If you have a lab PSU with current limiting, turn the current way down and test it
No chance of popped transistors then
 
AndyD273
AndyD273
I'd love to have something like that...
 
Giskard42
Giskard42
You can get a cheap Korad for <100 cad
Even something like this would be tremendously helpful: amazon.ca/dp/B00WEBJRE8?psc=1
@AndyD273 I've GTG for a few minutes, but tell me how it went!
 
AndyD273
AndyD273
Huh, I tried switching to the US amazon, and the canadian version is much prettier, with that case and all...
 
Giskard42
Giskard42
@AndyD273 If you're in the US there are probably better ones you can get
 
Telmo Marques
Telmo Marques
Hi everyone! Can I use a Choke instead of an Inductor?
Information on the web varies... some sources say they are the same things, other not so much...
Also, first time using the chat! Sorry if this is not the correct place to make these questions
 
W5VO
W5VO
6:30 PM
@TelmoMarques Ask a broad enough question, and you get the classic engineering response: "It depends"
 
Telmo Marques
Telmo Marques
@W5VO, fair enough :) I'm trying to build a reactive dummy load
The schematic is the same I've posted in this question
10
Q: Issue with parallel resistors in LTSpice

Telmo MarquesI'm trying to build a reactive load box, so I can run a guitar amplifier without a speaker. It's basically a device that simulates the impedance/frequency curve of a relatively high power speaker. I need 4Ω nominal impedance, but my local electronics store does not sell 4Ω 100W resistors, so I t...

resistors parallel ltspice
But I don't know which type of inductor to get
I believe the inductors are there to work as a filter. The choke definition fits that purpose, but I'd like to be sure.
 
Alice Ryhl
Alice Ryhl
7:09 PM
If I connected a non-inverting integrator and an inverting integrator in a loop, would I get a sine wave?
 
Tony Stewart
Tony Stewart
7:54 PM
@AliceRyhl only steady sine wave under certain conditions eg perfect unity loop gain at -180 phase shift.
 
AndyD273
AndyD273
Is there anywhere that would help me learn how to read a data sheet? I'm looking at this one: hebeiltd.com.cn/peltier.datasheet/TEC1-12706.pdf and I'm trying to figure out the performance curves, but I think I'm missing some basic concepts
 
Alice Ryhl
Alice Ryhl
I mean it sounds like it would be nontrivial to set the initial conditions to get the amplitude one wants, and it sounds like something where the sine wave would decrease in amplitude over time
 
AndyD273
AndyD273
Specifically, I'm trying to figure out what voltage/amps are needed to get a temperature of 40C out of it
But I don't know if that's what the performance curves even show
It kinda looks like I could do 1.5A at 5V... but I could be wrong
 
PlasmaHH
PlasmaHH
you dont get "40C out of it". if you want a temperature you need a control loop
 
Alice Ryhl
Alice Ryhl
do any of you know of a similar configuration to the integrator loop, that generates a sine wave
without being difficult to set the initial conditions
 
AndyD273
AndyD273
8:01 PM
@PlasmaHH From some of the stuff I'm seeing, it looks like I could under drive it and get a lower heat output. is that wrong?
3
Q: Are peltier devices linear?

cksa361Say I've got a peltier block that is rated at 15.2V and 6A, providing 90W of cooling power to one side. Will I be able to operate it at 5V, and will it draw 2A, providing 30W of cooling power? EDIT: sorry, error above. :P brain was not working. 5V and 2A would indeed be 10W instead of 30W above.

peltier
 
PlasmaHH
PlasmaHH
@AndyD273 it isn't, although a peltier isn't really the proper tool if you only do heating
 
AndyD273
AndyD273
@PlasmaHH It would be cool to have the option to use a DPDT switch to reverse current and cool instead. @Giskard42 suggested a Sand Resistor for heat, which might be fine too (more compact than needing a heat sink for cooling purposes) but since this is for experimentation I'm hoping to try out the peltier.
So is is possible to figure out what kind of current I need (or get close anyway) from looking at that data sheet? Or is that not how that works?
I'm a beginner, sorry :)
 
PlasmaHH
PlasmaHH
You need a heatsink then, which makes it even less ideal for heating
the "kind" of current you need depensds on what you want to do
 
AndyD273
AndyD273
What I'd like to do is get a semi consistant temp of around 40C (104F) since that's warm without risk of burning skin.
This in reverse:
 
PlasmaHH
PlasmaHH
I already told you what you need
 
AndyD273
AndyD273
8:15 PM
My boss is part eskimo, and so anything above sub arctic is too warm for him. I'm hoping to make a personal heater that will keep my fingers (and the rest of me) warm...
@PlasmaHH The heatsink fan wouldn't need to be on in cooling mode.
 
PlasmaHH
PlasmaHH
@AndyD273 wrong, it is specifically for cooling
 
AndyD273
AndyD273
Er, heating.
 
PlasmaHH
PlasmaHH
then its not just using a switch but reassembling things, then it might be better to have two devices, one for heating, one for cooling. faster to switch
 
AndyD273
AndyD273
Hmm... if you chill a heatsink, does it absorb heat from the air faster than if you just had the chiller without the heatsink?
 
PlasmaHH
PlasmaHH
eh, what?
 
AndyD273
AndyD273
8:26 PM
IIRC, heat flows from hot to cold. if the air is warmer than the heatsink, will it warm up faster than if you had a similarly cold object without the heatsink?
Like, take two bricks and throw them in the freezer, then pull them out and put them on a table. One has a big heatsink on it. Which warms up first?
 
PlasmaHH
PlasmaHH
the one with the greater surface obviously
 
AndyD273
AndyD273
So putting a heatsink on the cold side of a peltier might not reduce efficiency that much...
Sorry, random end of week thoughts
 
PlasmaHH
PlasmaHH
it actually increases it
 
AndyD273
AndyD273
So I'm not really understanding your comment it needing to be reassembled to switch from heating to cooling and back. You can just reverse polarity to change with side gets hot and which gets cold
If you have it so the side toward the skin gets cold, the heatsink would get rid of the excess heat, and if you want to warm the skin side the heatsink won't hurt anything...
I'll probably have to get some kind of variable PSU so I can play around with different voltages, see what will work best
 
PlasmaHH
PlasmaHH
8:42 PM
I understood you said you removed the heatsink when its in heating mode
 
AndyD273
AndyD273
I may have. I'm tired. More just not have the fan blowing, if it needs a fan.
Like the one in that video
At low temperatures a heatsink without a fan might be fine? I dunno. I guess I'd have to run it for a while and see how warm it got
 
PlasmaHH
PlasmaHH
again, you need a control loop
 
AndyD273
AndyD273
@PlasmaHH Ok. I was thinking of using a thermister and a transistor to turn off the current if it got to warm
 
PlasmaHH
PlasmaHH
you want something reacting quickly
 
Giskard42
Giskard42
I mean, it's not going to be a perfectly tuned pid loop, but it'll be pretty okay for just a hand warmer, no?
 
AndyD273
AndyD273
8:52 PM
I've read that they shouldn't be power cycled much, so I was hoping to run at a lower current to reduce the amount of heating needed
 
PlasmaHH
PlasmaHH
if it reacts quickly enough....
constant current buck converter is how I drive them
 
AndyD273
AndyD273
I was just looking at those based on @Giskard42's suggestion
 
PlasmaHH
PlasmaHH
the peltier thing reminds me of that I still need to do the lapping for the heatsink that I wanted to attach to mine...
 
AndyD273
AndyD273
Is a constant current buck converter different than a normal buck converter?
 
PlasmaHH
PlasmaHH
the normal buck converter is designed to provide a constant voltag
 
AndyD273
AndyD273
9:05 PM
The ones I was looking at appear to be adjustable, to provide lower voltages if needed, since I can't afford a lab PSU...
 
PlasmaHH
PlasmaHH
the question si what you can adjust, voltage or current
 
AndyD273
AndyD273
This one looks like it might do both voltage and amperage?
 
PlasmaHH
PlasmaHH
9:39 PM
anyone has a good idea to put ~30 18650 cells in parallel with some quick self built thing (instead of buying proper holders and/or welding contacts to them)?
 
AndyD273
AndyD273
@PlasmaHH two boards with contacts fixed at precise intervals that can be clamped tight to maintain contact?
Seems like parallel would be easier than series... since the contacts could be connected
 
PlasmaHH
PlasmaHH
if by boards you mean pcb, then thats hardly something I can do today... if you mean wood with something as contacts then maybe... I dont know if I have some, I would have to go into the basement to check
 
AndyD273
AndyD273
I was kinda picturing wood. It's the first think that comes to mind anyway. Since all the - are connected and all the + are connected, you just need some conductor that runs to each terminal... Maybe a box just big enough for all 30 to stand upright, with a metal bottom, and then another metal sheet to lay across the top... But I'm a newbie at this so I could be way off.
Alternately, metal thumb tacks with wire wound around them before being pushed into the board. Then you just have to messure the distance center to center when the batteries are side by side.
just need ones with low resistance
Well, thanks for the advice all!
 
Skyler
Skyler
10:41 PM
@Shalvenay so I got those LED drivers, one little issue
 
Shalvenay
Shalvenay
@Skyler yeah, they are a bit on the small side :) nothing a sot-23 breakout board can't fix though
 
KyranF
KyranF
a 3-pin LED driver? What does it do? What's the part number?
 
Shalvenay
Shalvenay
 
KyranF
KyranF
10:56 PM
wow, that's neat
 
Giskard42
Giskard42
@Shalvenay Breakout board? Nay, just forgo daily coffee and use lots of flux :)
 
KyranF
KyranF
standard pitch pin headers soldered on in strange angles will work fine too!
 
Giskard42
Giskard42
@KyranF The only acceptable solution is clearly flywires
Oh! Oh! Wire-wrapping!
 

« first day (2275 days earlier)      last day (2658 days later) »