|
I've got a couple of 34401A's that I'm currently using to measure the output of the EDC CR103 that I'm currently using as my voltage source. Do you think that'll do?
|
#
?
Jun 29, 2012 01:40
|
|
|
|
| # ? Jun 11, 2024 18:46 |
|
|
Otto Skorzeny posted:Accuracy doesn't have to be that great, I just want something real, real stable - 4 uvolts RMS of ripple would be good, 1 would be ideal. I'm using this as a reference to see if a vendor is full of poo poo (or perhaps more accurately how full of poo poo they are) in their claim that their 20-bit sigma delta ADC is capable of an ENOB of 19.2 bits at 10 updates/second. Really I only want 17 bits and could live with about 16.3, but right now I'm seeing 18 bits near zero and 14.1 bits near full scale (1.024v), degrading just about linearly through the range, even after applying all sorts of filtering and massaging. I suspect that the ADC's internal VREF is the problem, but before addressing that I want to eliminate the signal that I'm feeding the thing as a source of the poor precision. What's your noise integration BW? Hittite has some LDOs with very low noise: http://www.hittite.com/products/view.html/view/HMC976LP3E You should be able to drop the voltage with a resistor divider. Use the smallest resistors you can to minimize the thermal noise contribution. Another way to reduce the noise even further is to drive the base of a NPN with an RC filter from a voltage reference, and the collector with the input power rail. The emitter is the output. This won't give you a regulated voltage, nor will it be temperature compensated, but it can reduce the noise on the reference by quite a bit (depends on how noisy the reference actually is). It also has the side effect of providing a diode drop from your reference voltage, which could be desirable if you want a lower voltage.
|
|
#
?
Jun 29, 2012 01:55
|
|
|
Otto Skorzeny posted:I've got a couple of 34401A's that I'm currently using to measure the output of the EDC CR103 that I'm currently using as my voltage source. Do you think that'll do? Probably acceptable for measurements made in one session. The 34401A drift specs are a bit wide if you want microvolt repeatability across days.
|
|
#
?
Jun 29, 2012 03:51
|
|
|
Does anyone here have experience with Antec/Igloo FPGAs? For a possible future project I'll need some physically small FPGAs (less than 1cm square), but Altera doesn't seem to make anything in that range (and I'm not touching Xilinx). I'm not exactly sure what I need the thing to do, but it won't require much DSP or IP peripherals. The thing is Antec's small stuff seems to have way less logic elements than even low end cyclone FPGAs. And the logic elements in Igloo stuff (a VersaTile is a 3 input LUT or a flip flop) are less capable than the Cyclone elements (a 4 input LUT and a flip flop). So I'm wondering if the Igloos are even suitable for anything besides support roles. edit: oh, and it looks like Lattice Semiconductors makes some small stuff as well. Any insight on them would be appreciated as well. ANIME AKBAR fucked around with this message at 00:14 on Jun 30, 2012 |
|
#
?
Jun 29, 2012 23:43
|
|
|
I am putting together a mic preamp from a sparkfun circuit board, and the guy in the electronics shop where I got my components sold me 220 mOhm resistors (Red Red violet Red) instead of 220k Ohm resistors (Red Red Yellow Gold) I take it that these aren't interchangeable? http://www.sparkfun.com/products/8872 this is the board.
|
|
#
?
Jun 30, 2012 01:09
|
|
|
Brekelefuw posted:I am putting together a mic preamp from a sparkfun circuit board, and the guy in the electronics shop where I got my components sold me 220 mOhm resistors (Red Red violet Red) instead of 220k Ohm resistors (Red Red Yellow Gold) Almost certainly not. Also, you meant capital 'M'.
|
|
#
?
Jun 30, 2012 01:13
|
|
|
Brekelefuw posted:I am putting together a mic preamp from a sparkfun circuit board, and the guy in the electronics shop where I got my components sold me 220 mOhm resistors (Red Red violet Red) instead of 220k Ohm resistors (Red Red Yellow Gold) Is it "Red Red Violet Red" (4 bands) or "Red Red Violent" or "Red Red Red" or something? I have no idea what you mean. I'm not sure if I've even seen sub-1 ohm resistors in that formfactor before. Is violet-red a thing for 10^-2? quote:Almost certainly not. Also, you meant capital 'M'. Delta-Wye fucked around with this message at 01:32 on Jun 30, 2012 |
|
#
?
Jun 30, 2012 01:29
|
|
|
Sorry, I know nothing about this stuff. I am just learning as I go, and if I run in to more troubles I will talk to the two electronics guys I work with after the weekend. I guess I will stop putting this thing together until I can get the new resistor then. I do have soldering experience at least. I build a PAIA max Theremin kit about 8 years ago.
|
|
#
?
Jun 30, 2012 01:44
|
|
|
Brekelefuw posted:Sorry, I know nothing about this stuff. I am just learning as I go, and if I run in to more troubles I will talk to the two electronics guys I work with after the weekend. If you don't mind soldering/desoldering, you should just try it. Looking at the schematic, it shouldn't hurt anything; I don't think it will sound right but it could be an interesting experiment.
|
|
#
?
Jun 30, 2012 02:44
|
|
|
Delta-Wye posted:If you don't mind soldering/desoldering, you should just try it. Looking at the schematic, it shouldn't hurt anything; I don't think it will sound right but it could be an interesting experiment. I don't think the value 220K is especially critical there, but 220M is really huge, and the transistor probably won't be biased into conducting. If you have anything in the 100K-1M range that might work fine, depending on the transistor.
|
|
#
?
Jun 30, 2012 03:00
|
|
|
That looks to be true, though it won't be able to drive any sort of real load. Looks like you'll run into the upper rail and have a ton of distortion.
|
|
#
?
Jun 30, 2012 03:02
|
|
|
 I'm encouraging 
|
|
#
?
Jun 30, 2012 03:21
|
|
|
sixide posted:That looks to be true, though it won't be able to drive any sort of real load. Looks like you'll run into the upper rail and have a ton of distortion. It would help to short that LED, although it's not a Sparkfun project without an LED.
|
|
#
?
Jun 30, 2012 03:21
|
|
|
I have a 12V lead-acid battery and a photovoltaic panel that I hope to use to charge the battery to run some LEDs to light my back porch. Is there anything special I need to do in order to trickle charge the battery with the PV panel, or do I just connect it to the battery terminals? I know that overcharging Li-ion batteries is bad, bad news, but I know nothing about lead-acid batteries. If I do need some sort of circuit to cut off the PV panel when the battery reaches 12V, could someone give me some tips as to how to accomplish it? The only method I can think of right would be to use a resistor divider to get the battery voltage down to a max of 5V so it can be input to a microcontroller's ADC.
|
|
#
?
Jul 1, 2012 03:03
|
|
|
Otto Skorzeny posted:Accuracy doesn't have to be that great, I just want something real, real stable - 4 uvolts RMS of ripple would be good, 1 would be ideal. I'm using this as a reference to see if a vendor is full of poo poo (or perhaps more accurately how full of poo poo they are) in their claim that their 20-bit sigma delta ADC is capable of an ENOB of 19.2 bits at 10 updates/second. Really I only want 17 bits and could live with about 16.3, but right now I'm seeing 18 bits near zero and 14.1 bits near full scale (1.024v), degrading just about linearly through the range, even after applying all sorts of filtering and massaging. I suspect that the ADC's internal VREF is the problem, but before addressing that I want to eliminate the signal that I'm feeding the thing as a source of the poor precision. If it's the PSoC3/5, you won't get better than 17-ish bits of accuracy with the internal Vref, it's only 0.1% and the tempco is a little loose too. That's with a PSoC3, the current PSoC5 is specced at 1%. You're only ever going to get 19.2ENOB with careful setup and an external reference.
|
|
#
?
Jul 1, 2012 14:31
|
|
|
ANIME AKBAR posted:Does anyone here have experience with Antec/Igloo FPGAs? For a possible future project I'll need some physically small FPGAs (less than 1cm square), but Altera doesn't seem to make anything in that range (and I'm not touching Xilinx). I'm not exactly sure what I need the thing to do, but it won't require much DSP or IP peripherals. I assume you mean Actel/Microsemi. Those are the guys that make Igloo. I don't have any experience with them, but I think they're a pretty big player in rad-hardened stuff for space and mil-aero applications. That may carry over to their non-rad hardened stuff too, especially since they have to use large geometry for the rad-hard parts, and it would be expensive to keep several fabs open for smaller process technologies. Do you have any ideas as to what you need to do with the FPGA? You're going to have a hard time getting a lot of logic in a really small package, just due to the physical die size and the amount of pins you need for power. The PCB and manufacturing costs will also go up when you move to the really tiny parts (like a dense CSP with 0.4mm BGA) if you need via-in-pad or microvias. The other thing to consider is the development tool set. Since Altera and Xilinx are the biggest FPGA vendors, their tools and support will be far better than one of the second tier companies. I've used Lattice parts before, and developing for them kinda sucks compared to Altera/Xilinx. Granted that was a few years ago, so maybe they've gotten better since then. They do make small parts though. Any particular reason you don't want to use Xilinx?
|
|
#
?
Jul 1, 2012 16:09
|
|
|
Dielectric posted:If it's the PSoC3/5, you won't get better than 17-ish bits of accuracy with the internal Vref, it's only 0.1% and the tempco is a little loose too. That's with a PSoC3, the current PSoC5 is specced at 1%. You're only ever going to get 19.2ENOB with careful setup and an external reference.  this is about what we suspected but is rather different than the tune the apps engineer was singing. I hope it was just miscommunication (although we showed him our schematic and our code).
|
|
#
?
Jul 1, 2012 16:31
|
|
|
SnoPuppy posted:Do you have any ideas as to what you need to do with the FPGA? quote:You're going to have a hard time getting a lot of logic in a really small package, just due to the physical die size and the amount of pins you need for power. quote:The other thing to consider is the development tool set. If Altera had some parts that fit the bill, I'd probably go for them, but I don't see any (and their requirement for external configuration memory is a drag). I'm trying to do some deep investigation now because I don't want to have to switch platforms midway through. This is probably the first design I'll be trying to pitch for a major industrial application, so I want to choose the best part for the application, not the one that I just happen to be the most comfortable with.
|
|
#
?
Jul 1, 2012 17:02
|
|
|
Welp, I fried my Spartan 3E FPGA board tonight. A 12V wire touched some pin on the board while I was moving it, and it was dead.  That's what I get for being careless and messy with my workbench. Thankfully, it wasn't doing a whole lot for my project, just some basic logic. Now, if I lost my Spartan 6 board, I would be hurting.
|
|
#
?
Jul 2, 2012 09:18
|
|
|
I'm looking at buying a new soldering iron because my orange Weller from Sears has 3 tips and they're all too big for what I need to do. Im looking to solder some tiny 7135 chips to each other in parallel and the pins are drat tiny. Can anyone help me out with a recommendation of a good soldering iron for tiny electronic things? Preferably under 100 USD.
|
|
#
?
Jul 2, 2012 15:50
|
|
|
ANIME AKBAR posted:Yes, design and development tools are a huge deal. I'm okay with quartus, but last time I tried Xilinx's stuff it was awful. poo poo just didn't work and nobody could ever figure out why. Simply getting our license to function took over a week. Almost no support whatsoever from Xilinx. I've never met a single engineer at my university who preferred Xilinx to Altera (mostly because of the design tools). I'd vote for Altera, I love working with Stratix II/V/VI and Cyclone IV/V + Quartus; the GUI has gotten a lot better, and Quartus is also very scriptable; we have a Stratix project that I don't think we've ever used the GUI for, just some TCL scripts for housekeeping and setting up pin-outs for various board revisions. Looks like it may not be possible for you to use though
|
|
#
?
Jul 2, 2012 15:54
|
|
|
foundtomorrow posted:I'm looking at buying a new soldering iron because my orange Weller from Sears has 3 tips and they're all too big for what I need to do. I've really been quite happy with: http://www.adafruit.com/products/303 The only annoyance is the fact that it lacks a power LED (the LED only indicates whether the iron is at heating up). But that's minor, since it's a pretty solid choice.
|
|
#
?
Jul 2, 2012 16:24
|
|
|
That's eevblogs iron of choice. A buddy of mine picked one up, and I've had the chance to use it. It's really very nice.
|
|
#
?
Jul 2, 2012 16:36
|
|
|
I've been using this station for the past 4 years now. Absolutely stellar, as good as some stations that cost twice as much. http://www.mpja.com/Solder-Station-with-Digital-Temperature-Control-Display-302A/productinfo/15141+TL/
|
|
#
?
Jul 2, 2012 20:30
|
|
|
Minor update on the ADC signal stability stuff. With improved software (a really, really aggressive IIR filter plus some other jazz) and using a laboratory standard cell as both the analog Vref and the input on a PSoC 3, I managed to get to about 85k counts at full scale. I've got some ultra low noise LDOs coming in tomorrow morning to try, and some coefficients and minor other ideas to poke around with on the software end. Also, I think I'm going to badger Cypress with regards to getting some eval silicon for the PSoC 5 LP, since there's no way to test a bunch of our code on the PSoC 3.
|
|
#
?
Jul 3, 2012 00:04
|
|
|
movax posted:I'd vote for Altera, I love working with Stratix II/V/VI and Cyclone IV/V + Quartus; the GUI has gotten a lot better, and Quartus is also very scriptable; we have a Stratix project that I don't think we've ever used the GUI for, just some TCL scripts for housekeeping and setting up pin-outs for various board revisions.
|
|
#
?
Jul 3, 2012 00:33
|
|
|
Anybody rockin' a Rigol scope? I picked up a used 2-channel 60MHz / 16-channel digital recently, and I was wondering who else has one?
|
|
#
?
Jul 3, 2012 06:41
|
|
|
Otto Skorzeny posted:Minor update on the ADC signal stability stuff. With improved software (a really, really aggressive IIR filter plus some other jazz) and using a laboratory standard cell as both the analog Vref and the input on a PSoC 3, I managed to get to about 85k counts at full scale. I've got some ultra low noise LDOs coming in tomorrow morning to try, and some coefficients and minor other ideas to poke around with on the software end. Couldn't you just add a shitload of capacitance to whatever output you want and call it good?
|
|
#
?
Jul 3, 2012 06:52
|
|
|
Otto Skorzeny posted:Minor update on the ADC signal stability stuff. With improved software (a really, really aggressive IIR filter plus some other jazz) and using a laboratory standard cell as both the analog Vref and the input on a PSoC 3, I managed to get to about 85k counts at full scale. I've got some ultra low noise LDOs coming in tomorrow morning to try, and some coefficients and minor other ideas to poke around with on the software end. Awesome; I haven't done anything with a ADC in awhile, but I forgot how utterly frustrating it could be. ANIME AKBAR posted:That's good to hear. Altera actually does list a Cyclone IV device in a 8x8 mbga package (the EP4CE15M8), but I'm unable to find it for sale anywhere. In fact I'm not even sure it actually exists, because searching for it on google yields no results except those same Altera documents. Maybe they cancelled it, or it's just not "out" yet; we were intending to target a Cyclone IV GX15, which was advertised/listed right below the GX30, but it actually became available for retail and not just ES like six months later. So now we have some TCL scripts to auto-generate/manage pin-outs/timing for both the GX15 and GX30  Re: soldering iron talk, I have a MX-500 from Metcal, and every job I've ever had used only Metcals. I got mine for $150 "broken" and replaced a few caps in the base unit/power supply. Have only had to buy new tips so far.
|
|
#
?
Jul 3, 2012 15:28
|
|
|
ANIME AKBAR posted:Unfortunately that's not really clear. I'm going to make an RF amp which must be controlled via fiber optic, and the FPGA is going to handle that interface. But I don't really know what kind of communication scheme will be used (it may or may not be under my control). That sucks about your Xilinx experience, I actually prefer using them to Altera. Of course, it may just be that I've used their parts/tools for so long that I've gotten use to all the quirks. Since you don't know what you have to have in your design, you should spec in the largest part you can fit. It's easy to go down in size, but much harder to go bigger when you've finished layout of the board. You might want to play around with the Lattice tools to see how you like them - I remember things being painful, but that was years ago. Hopefully they're much better now. You mentioned needing a fiber interface - do you need high speed serial transceivers for a XAUI link?
|
|
#
?
Jul 3, 2012 17:00
|
|
|
I am daydreaming of an electronics project, that may never come to pass. One thing I would need is a circuit that could pulse a set number of times. What I need (for my feeble mind) is six different input pins. When each of these is pulsed once, each sends a set number of pulses out of an output pin. pin1 = 1 pulse pin2 = 2 pulses pin3 = 4 pulses pin4 = 8 pulse pin5 = 16 pulses pin6 = 32 pulses if it can do 7/64 and 8/128 that would be wonderful, but I can get by with just the first six. the input frequency would be low, only 50-100 htz. The pulse duration would be at least 50ms. And I don't want an arduino or raspberry or any other min-OS loving things up along the way, I need plain TTL here. I never messed with electronics much and never got into digital stuff. Is there a chip that will do this already? And it is cheap? I remember that the 7400 series chips had plain logic gates, but I'd prefer to avoid wiring the whole thing up myself....
|
|
#
?
Jul 3, 2012 23:15
|
|
|
edmund745 posted:I am daydreaming of an electronics project, that may never come to pass. One thing I would need is a circuit that could pulse a set number of times. "I want this specific functionality. But I don't want to use a microprocessor. Or logic gates..." I would knock it out in a 50-line C program on whatever microcontroller I had around, but that's because I'm too lazy to wire up a bunch of decade counters and a time source (555 or something). TTL logic gates would work too but you'd need a few to get the behavior you're wanting. What is this for?
|
|
#
?
Jul 3, 2012 23:28
|
|
|
It sounds like a perfect fit for a CPLD or small FPGA.
|
|
#
?
Jul 3, 2012 23:38
|
|
|
edmund745 posted:Is there a chip that will do this already? And it is cheap? A CPLD. Any chip that would have such a specific function is going to be more expensive because of how specialized it is which translates to less demand.
|
|
#
?
Jul 4, 2012 00:00
|
|
|
A CPLD or a microcontroller would be the way to go. A CPLD will be able to give you better accuracy, and I'd say if you've never worked with either or done any C programming, they're more-or-less equivalent in terms of difficulty. Maybe microcontrollers have a slight ease-of-use advantage because you can get them in DIP packages.
|
|
#
?
Jul 4, 2012 00:32
|
|
|
Delta-Wye posted:"I want this specific functionality. But I don't want to use a microprocessor.  quote:What is this for? After looking I see that there is a mega-arduino that should (?) work (54 pin-outs). I didn't want it to have to do any complex processing, but it can be 'permanently' connected and if all it is doing is translating between a PC and the servos then that sounds possible. The arduino wouldn't need to store huge amounts of data or do much of any logic itself, and it runs as fast as a USB-1 connection would. ------- I was pondering building a stepper driver that worked off the PC screen, which was what the above request was for. It could be extremely simple electronically, but it would also be extremely slow run directly off the screen rate. A screen only refreshes at 60 hz or so. Most steppers are 200 pulses/turn and 4 windings, so to turn the stepper around just once would take ~13 seconds. With the above pulsing circuit the max pulse rate this way would jump to 32x60, or 1920 pulses/sec or .4~seconds, a lot more acceptable. I still DO like this idea, really. But if it takes a separate microprocessor to do it, then one might as well use the separate microprocessor directly anyway I suppose.
|
|
#
?
Jul 4, 2012 01:02
|
|
|
edmund745 posted:
I think a dsPIC suffixed with 'MCxxx' would be perfect; 16-bit, program in C, and bitch-tons of PWM outputs with very flexible configuration. Naturally the dsPIC will have CMOS-level outputs, but it is a simple matter to level-shift. A CPLD/FPGA would also kick-rear end at this, but they are not so amenable for hobbyists; at best, I think you would get a TQFP, at worst BGA.
|
|
#
?
Jul 4, 2012 01:06
|
|
|
edmund745 posted:
I'm having a hard time following why your drive rate is tied directly to your screen refresh. Any reason why you can't use 12 stand-alone stepper controllers? Cost?
|
|
#
?
Jul 4, 2012 01:29
|
|
|
SnoPuppy posted:That sucks about your Xilinx experience, I actually prefer using them to Altera. Of course, it may just be that I've used their parts/tools for so long that I've gotten use to all the quirks. quote:Since you don't know what you have to have in your design, you should spec in the largest part you can fit. It's easy to go down in size, but much harder to go bigger when you've finished layout of the board. quote:You might want to play around with the Lattice tools to see how you like them - I remember things being painful, but that was years ago. Hopefully they're much better now. Also I found out that although Lattice's iCE40 stuff has nonvolatile configuration memory, it's just a PROM and can only be programmed once. So I'd have to use an external flash anyways. And Actel's Igloo stuff does have true FLASH configuration memory, but it's only rated for 500 write cycles. Definitely not enough for a newbie like me, so again it seems I'll need some external configuration no matter what (to start at least). quote:You mentioned needing a fiber interface - do you need high speed serial transceivers for a XAUI link? But that brings up another issue, which is that my research advisor has it in his head that each of these things are going to cost less than $40 in parts. I don't think he realized that even low end fiber receivers are all $15 or more, and we'll need two per board. ANIME AKBAR fucked around with this message at 03:06 on Jul 4, 2012 |
|
#
?
Jul 4, 2012 03:01
|
|
|
|
| # ? Jun 11, 2024 18:46 |
|
|
Delta-Wye posted:I'm having a hard time following why your drive rate is tied directly to your screen refresh. This would get around the parallel-port/USB issues totally, but the fastest you could turn each photocell on would be (!) 30 times per second (not even sixty times, as I put above). But if one screen pulse could represent up to 32 stepper pulses it couldn't be nearly as bad. quote:Any reason why you can't use 12 stand-alone stepper controllers? Cost?
|
|
#
?
Jul 4, 2012 03:16
|
|