This board already has 48 (!) SSRs in the channel switching paths, not to mention 9 real relays for the BIST switching. But a 6:1 mux with ssr's would take more room and more port pins than I have.
We're using a Claire (alternate NEC) part that's about 85 cents in moderate quantity.
I agree, ssr's are very handy, even in situations where "relay" is an improbable choice. Think of them as isolated-base transistors.
John
I've been kind of following this thread, and I have a question: did I miss the part that said, "Well, yeah, the 21:1 divider protects the input "for the big stuff"" without noticing that the input of the 21:1 divider and the input of the 4:1 divider both follow the same signal?
IOW, what have you got protecting the 4:1 input from the 25Vpp that results from dividing 100V by 4?
Thanks, Rich
I find it hard to believe that a peer review involving more than say, ... 2, people would have missed that.
That's what the whole thread is about. I was assuming that the HC4051's esd diodes would catch the small current resulting from driving the 4:1 divider with 50 volts peak (only 12 volts unloaded) but as Jim figured out, even that 0.6 volts past the rails turns on some fets internal to the switch.
John
Exactly two people were involved in the design. It was done quickly, in the midst of a more serious crisis, and we knew we were rushing it and taking chances. The rev A boards, the first etch, work perfectly and are deliverable - Linux drivers are being tested at a major customer site already - with one jumper and one kluged zener diode.
Do your peer reviews work better than that?
John
My most recent bug (alas, undetected at peer review):
It was a PIC-based USB-interfaced display controller. I was doing the hardware design and most of the software, with one complicated routine provided by a software writer who had done similar functions on other projects.
I worked through most of the Holiday shutdown (with the promise of a makeup week off in April, when I had a week-long family event planned), over the next two weekends, and late Monday night, finally getting a display in the shipping box just in time for the last FedEx drop to our customer (who was having a demo for _his_ customer the next day). The hardware worked ok, but the software running on the test PC occasionally got the answer wrong for some of the data coming over the USB link.
The next morning, I looked over the complete PIC listing trying to find the problem, and found out that the routine written by the software guy that I had plugged in to my code was sending out the USB message it had encoded at the end of the routine. Unfortunately, I was also sending out the message right after the routine returned. I went running off to my boss to tell him I had a fix, only to find out that since we had succesfully made the shipment the night before, most of the team was being laid off. Right now. Pack your office.
I'll never make _that_ mistale again.
Richard,
Hopefully your boss delivered his news before you delivered yours.
Have something to add? Share your thoughts — no account required.
Ask the community — no account required