You need to use EasyEDA editor to create some projects before publishing
New User seeking feedback on first Schematic/PCB
1531 22
Darryl Lawler 5 years ago
Hello! I was hoping I could get some initial feedback for my first attempt at reproducing a simple LED light bar design. I created the initial Schematic using generic LEDs and then substituted the Cree XP-E2 packages.  I am unsure of whether I have selected those components properly based on the number of DRC conflicts I generate. I am hoping that the simplicity of this circuit lends itself to few errors with the PCB layout.  I used an existing PCB I had on hand, as a guide for the copper areas of the 11 nets.  I believe the large copper areas are necessary for Power Integrity in this instance. [https://easyeda.com/darryl.lawler/Light-Curtain](https://easyeda.com/darryl.lawler/Light-Curtain) I admit that I have been rather ham handed in my attempt, so please feel free to nit pick.  I need the criticism. Thank you, in advance, for the assistance.  Looking forward to building my skillset. EDIT: I added attached images of the reference PCB to my project page.
Comments
andyfierman 5 years ago
@Darryl Lawler, There are a number of changes you need to think about. It is not a good idea to directly connect LEDs in parallel to a voltage source without some sort of current balancing resistor or other mechanism. This can lead to big differences in brightness across the paralleled LEDs and could lead to catastrophic destruction if one LED starts to draw more current than the others as it will heat up, it's froward voltage will drop so it will draw more current and heat up even more and so it will go into thermal runaway. It is not a good idea to run diodes in single series strings (even a single diode) without some sort of current limiting mechanism. Basically the forward current in a diode is exponentially proportional to the voltage across it so every step of XmV change in forward voltage (i.e. your power supply) will produce the same _ratio_ of Y change in forward current, increasing exponentially as the applied voltage increases. For a way to demonstrate the problem see: [https://easyeda\.com/andyfierman/LEDs\_must\_have\_series\_resistors\-OoGYgCK2k](https://easyeda.com/andyfierman/LEDs_must_have_series_resistors-OoGYgCK2k) At 1.5A, a resistor in series is impractical because of the power it will dissipate. If you are reproducing someone else's design then the original designer clearly does not understand these issues. The normal solution is to connect a string of LEDs in series and use some sort of dedicated step up converter to drive them with a constant current (which may be adjustable depending on the driver chip. If you want more LEDs then you replicate the whole circuit. Examples of suitable chips: low voltage DC input: [https://lcsc\.com/product\-detail/LED\-Drivers\_ON\-Semicon\_NCP5006SNT1G\_ON\-Semicon\-ON\-NCP5006SNT1G\_C231824\.html](https://lcsc.com/product-detail/LED-Drivers_ON-Semicon_NCP5006SNT1G_ON-Semicon-ON-NCP5006SNT1G_C231824.html) [https://lcsc\.com/product\-detail/LED\-Drivers\_DIODES\_AL8806MP8\-13\_AL8806MP8\-13\_C130748\.html](https://lcsc.com/product-detail/LED-Drivers_DIODES_AL8806MP8-13_AL8806MP8-13_C130748.html) Mains input: [https://lcsc\.com/product\-detail/Others\_Diodes\-Incorporated\_AL1697\-20CS7\-13\_Diodes\-Incorporated\-AL1697\-20CS7\-13\_C326513\.html](https://lcsc.com/product-detail/Others_Diodes-Incorporated_AL1697-20CS7-13_Diodes-Incorporated-AL1697-20CS7-13_C326513.html) There are many suitable devices. I suggest you search somewhere like RS or Farnell as their search tools are much better than those at LCSC. Then see if you can find the chosen part  at LCSC. I'm not sure but it also looks like you may have tried to set the circuit up as a simulation (you included a DC voltage source from the simulation devices library). You cannot however simulate this circuit because there is no spice model available for the Cree diode.
Reply
Darryl Lawler 5 years ago
Thank you for the prompt reply.  Your feedback is certainly appreciated.  The circuit I'm replicating includes 9 other identical LED strips connected to a [Meanwell HLG-320H-30B](https://datasheet.octopart.com/HLG-320H-30B-Mean-Well-datasheet-10900476.pdf). There are 10 identical strips connected through a distribution PCB that puts a fuse in line with each of the 10 strips. The HLG-320H-30B has a constant current range of 15-30V DC.  It has a rated current of 10.7A.  It also has over current protection with the following note: > Constant current operation region is within 50% ~100% rated output voltage. This is the suitable operation region for LED related applications, but please > reconfirm special electrical requirements for some specific system design. I believe these parameters _might_ address the current balancing concerns?  Perhaps not.  I am a novice with circuit design, so please understand that I am only posing questions for the purposes of inquiry, not contradiction. Does the fact that these parallel connections are interconnected and therefore result in a mesh circuit more than a ladder or parallel circuit have any effect on balancing the currents? This: ![Screen Shot 2019-02-01 at 1.44.07 PM.png](//image.easyeda.com/pullimage/p2YMmJzcRi8oJ3yDKrZQgdQrLzDLfSTcNCDHQGkb.png) Versus this: ![Screen Shot 2019-02-01 at 1.45.21 PM.png](//image.easyeda.com/pullimage/zsoUwkQO9A1Ze6iuf1vcjCcaO1zZNuNWHKOIcP68.png) Additionally, the circuit I am replicating only has half the LED locations populated. i.e.: only 3 of each set of 6 LED locations actually have LEDs populating the solder pads. This can be observed in the photos I attached to the project page. Regarding the simulation, I have found SPICE simulation parameter text files for each of the LEDs I am hoping to implement. I think they should work to import that information into the Designer. Those files have the following content: > * Cree XLAMP XP-G2 LED > * Model valid for 100mA to 1500mA & Tj=25C > .MODEL XP-G2 D > > > > > > * IS=2.09328E-45 > * N=1.077479027 > * RS=0.260397666 > * XTI=30.8559152 > * EG=2.5000 > > > > > * Cree XLAMP XP-E2 White LED > * Model valid for 100mA to 1000mA > .MODEL XPEB_WHT D > > > > > > * IS=5.01824E-16 > * N=3.36413016 > * RS=0.244130334 > * XTI=40.92995615 > * IKF=0 > * EG=2.5000 > > > > > * Cree XLAMP XP-E2 Blue, Royal BLue LED > * Model valid for 100mA to 1000mA > .MODEL XPEB_BLRB D > > > > > > * IS=9.0628e-010 > * N=5.9196 > * RS=0.20196 > * XTI=53.34286 > * IKF=0 > * EG=2.5000 I apologize for not including more details like the power supply specs or other elements of the circuit in my original post. I certainly appreciate you offering the great advice and corrections you have. Thank you.
Reply
Darryl Lawler 5 years ago
An additional detail that may be of use:  These are high power commercial LEDs used in horticultural and architectural implementations.  Could the specifications of these LEDs be outside of the norm you are addressing with your concerns?  Again, this is a passive inquiry, not a contradiction of the facts you have presented.
Reply
Darryl Lawler 5 years ago
I am reading through your linked article, now.  I apologize if it addresses any of the items I mentioned in my reply.
Reply
andyfierman 5 years ago
Well done finding the spice model. I'd missed it: [https://www.cree.com/led-components/media/documents/XLamp-XPG2-Spice.txt](https://www.cree.com/led-components/media/documents/XLamp-XPG2-Spice.txt) See Table 7 on page 17 of: [https://www\.cree\.com/led\-components/media/documents/HD\_design\_guide\.pdf](https://www.cree.com/led-components/media/documents/HD_design_guide.pdf) ThAL8843 led driver from Diodes Inc. would be suitable for a DC input of between 4.5V up to 40V: [https://www.diodes.com/products/power-management/led-drivers/medium-voltage-dc-dc-led-drivers/part/AL8843](https://www.diodes.com/products/power-management/led-drivers/medium-voltage-dc-dc-led-drivers/part/AL8843) It can provide up to 60W of output power which for a single string of Cree LEDs running at a forward drop of typically 3.6V at 1.5A (p4 of [https://www\.mouser\.co\.uk/datasheet/2/90/ds\_XPE2\_Torch\-519361\.pdf](https://www.mouser.co.uk/datasheet/2/90/ds_XPE2_Torch-519361.pdf)) means that a string can be up to N LEDs in series where N is given by: N = NEAREST INTEGER BELOW(60W/(3.6W*1.5A) = 11. So for say, 33 LEDs, you need three off AL8843 chips each driving a series string of 11 LEDs. You can put fewer LEDs in a string so you could do 40 LEDs as 4 off drivers driving10 LEDs in their string. Unfortunately these devices do not have spice models available though I have done simulations with my own behavioural models for this type of device. They are also not available from LCSC. OnSemi do a range of parts: [https://www.onsemi.com/PowerSolutions/parametrics/16180/products](https://www.onsemi.com/PowerSolutions/parametrics/16180/products) as do many other manufacturers.
Reply
andyfierman 5 years ago
Sorry, missed a bracket: N = NEAREST INTEGER BELOW(60W/(3.6W*1.5A)) = 11.
Reply
Darryl Lawler 5 years ago
andy, Thank you for the thorough replies.  The depth of information and guidance is very appreciated.  I will work through some configurations utilizing the driver chips you suggest and see what I can figure out. A few quick questions: I'm going to be using different colors (the HE photo red of the XP-E, the royal blue and the cool white from the XP-G3), therefore there are different specifications for each LED.  Would you recommend segregating my colors so they have independent drivers? Or, are there options to mix the color of LEDs driven by a chip? In either case, would you have any recommendations for the following configuration: 2 royal blue [XP-G3](https://www.cree.com/led-components/products/xlamp-leds-discrete/xlamp-xp-g3)  ([Spice](https://www.cree.com/led-components/media/documents/XLampXPG3Spice.txt)) 19 neutral white min 90 CRI [XP-G3](https://www.cree.com/led-components/products/xlamp-leds-discrete/xlamp-xp-g3) ([Spice](https://www.cree.com/led-components/media/documents/XLampXPG3Spice.txt)) 9 HE photo red [XP-E](https://www.cree.com/led-components/products/xlamp-leds-discrete/xlamp-xp-e) ([Spice](https://www.cree.com/led-components/media/documents/XPE_SPICE.txt)) *note - the Cree LED parts have changed.  Apologies for the imprecision of my initial part selection. I was working from old notes. The model number text is a link to the product web page and the Spice text is a link to the respective Spice models. Orientation on the PCB (20mm x 560mm) would be as follows: WWWRRRWWBWWWWWWRRRWWWBWWRRRWWW± Again, thank you for your assistance. I am grateful for any help I can get and this is more than I could have hoped. I will get to work on this today.
Reply
andyfierman 5 years ago
Checked the perceived brightness of the different colours at the same drive current. All the LEDs in a string pass the same current through them so if you mix colours in the same string you must be sure that you don't end up with one colour appearing disproportionately brighter than the other. You can do it from the datasheet but it's complicated because you have to interpret the beam pattern graphs carefully. You might be better to just put some test strings together. If you have a bench supply that can source a 1.5A constant current that's easy or at least for testing purposes you could lash up a linear constant current source. The current setting resistor and pass transistor will get hot but for a test bench you can probably live with that if you get a suitable power resistor and transistor that can be bolted to a heatsink.
Reply
andyfierman 5 years ago
Don't you just love Autocorrupt? "**Check** the perceived brightness of the different colours at the same drive current."
Reply
Darryl Lawler 5 years ago
Thank you, again.  I'll get to my homework. ;-) Just to clarify, the Meanwell power supply ([Meanwell HLG-320H-30B](https://datasheet.octopart.com/HLG-320H-30B-Mean-Well-datasheet-10900476.pdf)) I mentioned in my extended reply above provides no constant current protection or any type of feature that would address the concerns we're solving with these LED driver chips?  The only way to address the variance in currents due to forward voltage behavior across the diodes is to use the LED driver chips?
Reply
Darryl Lawler 5 years ago
For additional reference material regarding my inquiry, page 7 of the Horiculture LED Portfolio Data sheet ([link](https://www.cree.com/led-components/media/documents/CreeXLampHorticultureFeatureSheet.pdf)) provides a parts list for their reference design.  They use a variation of the HLG-320H family of drivers without mention of the circuit design.  This is certainly not conclusive that their circuit does not include the LED driver chips you are suggesting, it just seems enough information for my intuition to pose the question of whether they are necessary with this specific type of driver.
Reply
andyfierman 5 years ago
"... whether they are necessary with this specific type of **driver**." Did you mean: "... whether they are necessary with this specific type of **LED**."
Reply
Darryl Lawler 5 years ago
This article ([link](https://www.ledsupply.com/blog/constant-current-led-drivers-vs-constant-voltage-led-drivers/)) seems to suggest that the power supply being a constant current power supply with a range of 5.35 - 10.7A, is intentionally selected to compensate for the variance in forward voltage demands and adjust its voltage accordingly.
Reply
Darryl Lawler 5 years ago
To answer your request for clarification: I am asking whether the LED driver chips are actually required given that the driver (aka "Power Supply) is a constant current source which has the specifications linked ( [Meanwell HLG-320H-30B](https://datasheet.octopart.com/HLG-320H-30B-Mean-Well-datasheet-10900476.pdf)) Does not the constant current circuitry within the [Meanwell HLG-320H-30B](https://datasheet.octopart.com/HLG-320H-30B-Mean-Well-datasheet-10900476.pdf) provide the current control you are suggesting the LED driver chips would provide?
Reply
andyfierman 5 years ago
The PSU you've identified is a 30A constant current source with an output voltage of from 15V to 30V (i.e. an output voltage compliance of 15V). With roughly 3.5V drop per LED that will drive up to 4 LEDs in series. As I've tried to explain above, you need to drive the LEDs in series (i.e. 1 LED in series with the next with none in parallel) with a constant current source and you need a constant current source per string. So if you have N LEDs with an average forward drop of Vf then your current source has to have a compliance of N*Vf. Your chosen LEDs have a max forward current of about 1.5A so if you connect the 4 or fewer of them in series to a constant current source of 30A you will destroy at least one, possibly all 4 of them. So your if your LEDs have a forward current rating of iF then your constant current source has also to have an output current of iF. If you try to connect more LEDs in series to it then the current will drop but not in a controlled way because the Meanwell PSU is being operated outside it's specified output voltage compliance range. If you try to connect more than one string in parallel then there is no mechanism to control how current goes down one string vs. any other. This all means that the Meanwell PSU isn't suitable. Essentially what you need is a DC voltage output PSU with suitable output short circuit protection and M constant current drivers where M is the number of strings.
Reply
andyfierman 5 years ago
My bad. Trying to answer technical questions late at night. Bad plan. With the Meanwell PSU you can drive 5 to 8 LEDs with a Vf of 3.5V because the output operates in constant current with an output voltage of between 15V and 30V. If you try to connect more than 8 LEDs in series to it then the current will drop but not in a controlled way because the Meanwell PSU is being operated outside it's specified output voltage compliance range. If you try to connect less than 5 LEDs in series to it then the output will probably drop into constant voltage operation because the Meanwell PSU is being operated outside it's specified output voltage compliance range. But as already explained, in the constant current source range of operation, it will force 30A through the LEDs. Which would be another bad plan.
Reply
andyfierman 5 years ago
A quick simulation of the simple 2 transistor linear constant current source/sink: [https://easyeda.com/andyfierman/linear-2-transistor-current-source](https://easyeda.com/andyfierman/linear-2-transistor-current-source)
Reply
Darryl Lawler 5 years ago
Thank you for your persistence in chipping away at my inexperience.  When you explained it, the obviousness was humbling.  I appreciate your patient attitude and solid advice.  Back to work.
Reply
MikeDB 5 years ago
Note that the Meanwell doesn't use the constant current source Andy has shown, but instead monitors the current and adjusts the voltage until it gets it right.  I've never used this particular unit, but I have seen others produce sizeable over-voltages if incorrectly loaded which is of course terminal for the LEDs.  Thus as Andy says, it really is important to ensure your load is within the specified band for the module, and always connected.
Reply
andyfierman 5 years ago
@MikeDB, Hmmm. "Meanwell" Kind of aptly named...
Reply
andyfierman 5 years ago
@darryl.lawler, This device: [https://www.diodes.com/products/power-management/led-drivers/medium-voltage-dc-dc-led-drivers/part/AL8861#tab-overview](https://www.diodes.com/products/power-management/led-drivers/medium-voltage-dc-dc-led-drivers/part/AL8861#tab-overview) from Diodes Inc would do a 1.5A string of up to about 7 LEDs from a 40V, 1.7A DC supply or 5 from a 30V, 1.7A DC supply. For M strings you would need a DC supply that would give you M*1.7A at whatever voltage you decide (i.e. the 30V to 40V). The AL8861 is a switch mode device so it can run at much higher efficiency than a linear constant current source and so will be much cooler. It's available in a SOT89-5 package which is easier to hand solder than some of the other, smaller packages with centre heatsink pads. Note however that although Diodes Inc claim to have a spice model for the device, the model is incomplete so it's useless. It might be possible to cobble something together from the (what looks to be a) complete model for the AL8806 or just put together a behavioural model but it's probably not worth the effort as it's not something I could do for free.
Reply
Darryl Lawler 5 years ago
Thank you for the suggestion.  I will keep it in mind and see what resources I may have to generate the behavioral model.
Reply
Login or Register to add a comment
goToTop
你现在访问的是EasyEDA海外版,建议访问速度更快的国内版 https://lceda.cn(需要重新注册)
如果需要转移工程请在个人中心 - 工程 - 工程高级设置 - 下载工程,下载后在https://lceda.cn/editor 打开保存即可。
有问题联系QQ 3001956291 不再提醒
svg-battery svg-battery-wifi svg-books svg-more svg-paste svg-pencil svg-plant svg-ruler svg-share svg-user svg-logo-cn svg-double-arrow -mockplus- -mockplus- -mockplus- -mockplus- -mockplus- -mockplus- -mockplus- -mockplus-@1x -mockplus-

Cookie Notice

Our website uses essential cookies to help us ensure that it is working as expected, and uses optional analytics cookies to offer you a better browsing experience. To find out more, read our Cookie Notice