--- title: "Link not working" url: "https://maker.wiznet.io/forum/11788" markdown_url: "https://maker.wiznet.io/forum/11788/md" type: "Forum topic" category: "Ethernet Chips" author: "Tinus81" created: "2016-04-11T21:42:09+09:00" last_activity: "2017-05-04T08:38:37+09:00" language: "en" views: 1909 replies: 10 source: "WIZnet Makers forum (https://maker.wiznet.io/forum/)" --- # Link not working ## Question Asked by Tinus81 on 2016-04-11 in Ethernet Chips. I’m having trouble getting a link with W5200. I have made two boards and both cannot get a link when connecting an ethernet cable. I’ve made the schematic following the W5200 reference schematic. I’m using the RB1-125BAG1A MagJack. I’ve checked all hardware connections and SPI works good. I’m able to set MAC and IP etc. I read the PSTATUS register as 0x07 every second, with or without cable connected. What could be wrong?! Here a part of my schematic: [![](https://maker.wiznet.io/forum_uploads/wiznet/original/1X/49c706749c2e735ade6a79cbde488dd43eb90a34.PNG) 1512×868 35 KB](https://maker.wiznet.io/forum_uploads/wiznet/original/1X/49c706749c2e735ade6a79cbde488dd43eb90a34.PNG) Regards, Martijn ## Replies ### Reply 1 by Eugeny, 2016-04-12 Martijn, you changed original circuit a little. 1. LED part. Why did you change it? W5200 senses its mode using LED outputs in input mode, are you sure putting P-MOSFET will not affect proper mode identification? 2. I assume you already saw reference schematic from [url][http://www.wiznet.co.kr/product-item/w5200/\[/url\]](http://www.wiznet.co.kr/product-item/w5200/%5B/url%5D). Please stick to it as much as possible. 3. pin 45 should be pulled down. If it is not, W5200 may power PHY down - probably this is what happens in your design. 4. CS, INT and RESET should have pull-ups to have proper levels when master does not (yet) drive the bus. 5. and finally, as you are able to read W5200’s registers, I assume you powered 1V8 properly. ### Reply 2 by Tinus81, 2016-04-12 Hi Eugeny, thanks for your reply. 1. I changed LED part because I need to drive more LEDs in the final product (which also will be PWM-ed). I desoldered the mosfets to test if that would work, but it doesn’t. I get no link. 2. Yes I did. For example, the power supply part was a direct copy of reference schematic. Now with the added resistors my schematic is equal to the reference schematic. 3. OK, I added a pull down resister to pin 45. 4. OK, I added pull up resistors to these pins. 5. Yes, I’m able to read and write registers. On the 1,8V pin I measure 2,09V, maybe it’s a little high? Even with the pull up and pull down resistors and removed mosfets I get no link. I’ve tested all w5200 pins, all are connected to the board ok. I’ve measured Auto Negotiation pulses on all the w5200 phy pins (17/18, 20/21), but only with an ethernet cable is connected. Without cable I don’t measure these pulses (shouldn’t the w5200 also send negotiation pulses?). I really have no clue where to look further or where the problem might occur ![:open_mouth:](https://emoji.discourse-cdn.com/twitter/open_mouth.png?v=12) Any ideas?? Thanks, Martijn ### Reply 3 by Eugeny, 2016-04-12 ![](https://avatars.discourse-cdn.com/v4/letter/t/977dab/48.png) Tinus81: > On the 1,8V pin I measure 2,09V, maybe it’s a little high? Not little, it is too high - 16% above nominal. Did you connect 1V8_OUT to 1V8D through ferrite bead with capacitors as recommended on reference circuit diagram (top right corner of the sheet)? My first guess that you have 2.09V simply because 1V8_OUT (regulator output) does not have any load. ### Reply 4 by Tinus81, 2016-04-12 [quote=“Eugeny”] Not little, it is too high - 16% above nominal. Did you connect 1V8_OUT to 1V8D through ferrite bead with capacitors as recommended on reference circuit diagram (top right corner of the sheet)? My first guess that you have 2.09V simply because 1V8_OUT (regulator output) does not have any load.[/quote] Yes I did connect 1V8 out through ferrite bead and all the capacitors are there. I’ll check the board tomorrow and measure the 1V8D pins. Thanks for your help so far! ### Reply 5 by Eugeny, 2016-04-12 Yes, you surely need to check your board. I hope chip did not fry. Probably you have some shortage, for example pulled up RSV (7) is shorted with 1V8 (8). ### Reply 6 by Tinus81, 2016-04-13 I’ve placed a new w5200 on the board, to be sure it’s not defect. I’ve checked for shortage, there are no shortages. When I power up now, the 1V8 measures 1V9, so that seems much better. However I still get no link… I’ve tried multiple network switches and cables by the way. Any other suggestion?? ### Reply 7 by Eugeny, 2016-04-13 Can you please post a photo of your board from both sides? Did you connect 3V3A to 3V3D through recommended circuit? You did not provide full circuit diagram related to W5200, thus I have to guess what is out of the image you posted. From what I see that digital part seems to work properly - XTAL, host-side I/O, PHYSTAT shows correct value, PHY is not in forced power down mode. Thus there’s most probably some issue at the TX/RX and/or magjack’s side. ### Reply 8 by Tinus81, 2016-04-13 Here are photo’s from top and bottom side: [![](https://maker.wiznet.io/forum_uploads/wiznet/original/1X/386c9c7b3971a3e930fb467b8c473eaf27498c4a.jpg) 1920×1424 963 KB](https://maker.wiznet.io/forum_uploads/wiznet/original/1X/386c9c7b3971a3e930fb467b8c473eaf27498c4a.jpg) [![](https://maker.wiznet.io/forum_uploads/wiznet/original/1X/e2f7eda8c8c234327cbf0a90ae50f2243169b1a8.jpg) 1920×1424 936 KB](https://maker.wiznet.io/forum_uploads/wiznet/original/1X/e2f7eda8c8c234327cbf0a90ae50f2243169b1a8.jpg) The resistors connected with rework wire are the extra pull ups and pull down. Also, here is full schematic: [Schematic.pdf](https://maker.wiznet.io/forum/legacyFile.asp?path=7EjQVdiOjTpAmfy2dZ3f13VsZGd.pdf) (55.9 KB) I really appreciate your help! ### Reply 9 by Eugeny, 2016-04-13 In my opinion your circuit diagram is correct. Layout can be improved and optimized, but in general it should work. Do R5 and L3 have any shortage between them? ### Reply 10 by Tinus81, 2016-04-13 You are my hero! ![:smiley:](https://emoji.discourse-cdn.com/twitter/smiley.png?v=12) The shortage should be there - but not with R5. It turns out that the text of R5 and C21 is reversed ![:blush:](https://emoji.discourse-cdn.com/twitter/blush.png?v=12) R5 is the bias resistor on pin 12. Thanks a lot!! --- Source: https://maker.wiznet.io/forum/11788