Subject: Re: Laser diode drivers for TGC From: Kevan Hashemi Date: Tue, 18 Jul 2006 12:04:20 -0400 To: Lorne Levinson CC: meir.shoa@weizmann.ac.il, Moshe Sidi , James Bensinger Dear Lorne, I'm back from vacation. Thanks for your patience. > We agree to remove L1 and L2. Splendid. > Note that the possible ground loop you describe is not possible... I see. Thank you for your explanation. > We could connect the shields of the incoming and outgoing cables together via a braid inside the box. Do you recommend this? No, that won't be necessary. > The laser board, cable and the driver box are assembled together > once, tested, and then delivered as a unit. They are not separated > after assembly. That sounds good. The lasers will be safe. > Bypass caps can be added. Which chips have highest priority? Each of the logic chips needs decoupling, but I think you need only one decoupling capacitor within 10 mm of each logic chip. > Is 0.1mF too high? By "0.1 mF" I assume you mean 0.1 uF, or 100 nF. That may be too high, depending upon the capacitor. I suggested 10 nF because almost all 10-nF ceramic capacitors will perform well at 1 GHz. If you look at this data sheet: http://www.avxcorp.com/docs/Catalogs/skycap.pdf you will see that they don't specify the frequency performance of the capacitors. It is possible that the 100-nF capacitor resonates at 100 MHz. So: as long as the decoupling capacitor resonant frequency is above 1 GHz, you will be okay. > We can also consider adding solid wire onto some of the PCB power/gnd > traces. With four or five decoupling capacitors, I think you will be safe. There is no need for added traces. But the added traces are unlikely to take care of the problem on their own. > The three times 47W is a mistake... I understand. Your proposed three 33 Ohm resistors will work fine. > Is the crucial cable length from USA15 to the drivers, or only from > the multiplexer? The root cable (from multiplexer to driver) is the critical cable, because it's the one carrying current for ten devices. > Note that the cable between diode and driver is always 50cm in > length. Okay. Thank you for your attention to my comments. I will be glad to test your next version in my lab, if you will send us one. Yours, Kevan -- Kevan Hashemi, Electrical Engineer Physics Department, Brandeis University http://www.bndhep.net/