I’ve got 35 years as an electronics design engineer doing microprocessors, hardware, and quite a bit of analog. In the late 1970s, I was involved with two different slewrate- limited applications.
One was control of the throttle and pitch of the props on patrol boats. The bridge could signal full ahead, and this condition would cause engine stalls. So, a slew-rate control was needed to bring the engine to full power and adjust the prop at the same time in the maximum rate that the engine could respond. (Good planning. /rap)
The other was back in the days when memory was expensive. I did a vectordrawn graphical display on a magnetic deflected CRT. A DAC08 was used to control the slew rates by setting the currents to the ring diodes. The system was to draw fonts at different sizes. (I did some of this too, when I was at Philbrick, designing digital-to-analog converters (DACs) for Loral in New York. /rap) A pair of 12-bit DACs set the character-start position. (The slew rate controlled movement. We used the core-sense amp for determining when slewing was done.) (I was designing 12-bit, low-glitch, fast-settling DACs for the major positions at ~6 MHz and 10-bit MDACs for drawing vectors. /rap)
A font-generation system had DAC control of the font size, six points to 128 points, with DAC control of font width and slew-rate control for angled vectors in the fonts. It all was done in analog, one feeding the next, and it worked like a champ. I loved those DAC08s and the slew-rate control. We needed fixed and stable supplies because of the way that I designed the system. I do not remember the part number but used a National laser-trimmed 10-V reference IC. (That was surely the LM168, which I got working. Some other poor turkey designed it, but I got it working and into production. /rap)
There were no pots or adjustments. With op-amp common-driven National threeterminal regulators, we got ±15 V to better than 10 mV with no adjustments, a stable temperature, and nothing that the customer could adjust to tweak the circuit out of spec. Now that’s doing analog.
Sounds good to me. I was on a Navy patrol boat off southern California. I looked at the wake, and it would veer and veer. I asked what the problem was, and the steersman said there was a lot of slack in the rudder linkages. I asked him, “Why don’t you just set one throttle a little bit slower, and the other one faster, and push against the rudder?” I don’t recall him taking my advice. But good engineers have to be smart and devious.
DEAR MR. PEASE,
I have a very simple question. We are in the process of designing a printed-circuit board (PCB) containing mixed analog and digital circuitry. The analog circuitry consists of balanced passive (RLC) filters. Would it make sense to remove ground planes under the filter sections to ensure there are no parasitic capacitive effects that could cause unbalance (I suspect that the ground planes will do more good than harm. I made a Heathkit FM receiver, and it was laid out carefully to reject 10.7-MHz noises and strays. But, it didn’t have any digital stuff on the same board. /rap) and convert common-mode signals to differential? (I’m guessing that differential is a better way, but I am not an expert on this. /rap) The frequency range of operation is up to 28 MHz.
This is up near the extreme end of my expertise. I am just guessing up there. I am very suspicious and skeptical of the concept of putting both these circuits on the same board. Do you want something cheap and cost-effective but doesn’t work? Or do you want it to work okay? Do you want it to work on the fifth try, or the second?
I suspect you would want to build the two circuits quite separately. If you’re lucky, putting the analog section in a separate shielded box will make it work okay. If you’re slightly unlucky, you’ll have to also shield the digital section in its own box. You would want to use your best layout engineer to get this. (I could be wrong.)
If you think you are fantastically lucky, you can later merge these two circuit boards into one board. And be prepared to mount copper fences and walls and screens between the analog and digital sections. How much guarding and shielding will you need?
“Are you feeling lucky, punk?” Working in places like that, I don’t feel very lucky. I could get each little PCB working, separately, 4 in. apart, but when moved close together, they can’t be made to work. There might be some engineers who would have confidence doing this, but not me.
What if your analog section has to have 35 dB of rejection of the digital noises? I suspect you have a chance. But if it’s 50 dB, it will probably take all those extreme efforts, and I still have no idea what can go wrong. I’m not an expert on L’s and C’s in passive filters.
And if you think I’d want to do some consulting on this problem, wrong! What I have just told you is going to be miraculous if you can get it to work.
Comments invited! [email protected] —or: r.a. Pease, 682 miramar avenue san Francisco, Ca 94112-1232