Nice story about your work at AEG. I can imagine like you said, if you send the wrong commands to switches in a power station, that hell breaks loose. Have seen some video's about a short with electric arc/fire in a high voltage cabin, where the operators where only wearing sandals , while the short / electric arc makes so much noise that you whould be deaf by the time the inferno is stopped. Grtz
From around 30:00 and on - you got the circuit function right, but the labels are swapped, IMO. DAC A output is course setting and DAC B is fine setting. R13, being lower in value, has much higher influence on the current flowing into the summing point, R14 has 100x lower contribution and therefore only makes minor (fine) adjustments to the output voltage. Since the R14 is 100x lower, it adds slightly less than 7 bits of resolution to the major 12 bits, making it slightly less than 19 bits of composite resolution. It would be interesting to see, whether they calibrated the contributions (that is, resistor values) once for all and they set the DAC according to that, or the composite DAC voltage value (on the opamp output) is checked and fine adjusted on the fly against the much more precise ADC.
I believe the DAC architecture is derived from the one described by Jim Williams in Linear Technology AN86. Like always, it is filled with plenty of other intriguing details.
Hi Alesaandro, thankyou for the info, haven't seen that AN-86 (A Standards Lab Grade 20-Bit DAC with 0.1ppm/°C Drift) from 2001 yet, yes that makes sense :-). Cheers, Konstantin
I love analog electronics so much, I wish we had it more often. Now I have one more interesting circuit to use
Hi Maksim, Please note that Fine & Coarse Labels need to be swapped in the Circuit I Shown, Typo :-)
Nice story about your work at AEG. I can imagine like you said, if you send the wrong commands to switches in a power station, that hell breaks loose.
Have seen some video's about a short with electric arc/fire in a high voltage cabin, where the operators where only wearing sandals , while the short / electric arc makes so much noise that you whould be deaf by the time the inferno is stopped. Grtz
Thankyou Bjorn!
80s were absolutely amazing
From around 30:00 and on - you got the circuit function right, but the labels are swapped, IMO.
DAC A output is course setting and DAC B is fine setting. R13, being lower in value, has much higher influence on the current flowing into the summing point, R14 has 100x lower contribution and therefore only makes minor (fine) adjustments to the output voltage. Since the R14 is 100x lower, it adds slightly less than 7 bits of resolution to the major 12 bits, making it slightly less than 19 bits of composite resolution.
It would be interesting to see, whether they calibrated the contributions (that is, resistor values) once for all and they set the DAC according to that, or the composite DAC voltage value (on the opamp output) is checked and fine adjusted on the fly against the much more precise ADC.
Hi, correct! Gain=R15/R14, so Labels are swapped.
Thankyou!
I believe the DAC architecture is derived from the one described by Jim Williams in Linear Technology AN86.
Like always, it is filled with plenty of other intriguing details.
Hi Alesaandro, thankyou for the info, haven't seen that AN-86 (A Standards Lab Grade 20-Bit DAC with 0.1ppm/°C Drift) from 2001 yet, yes that makes sense :-). Cheers, Konstantin
Maybe at that time 20bit DACs were way more expensive
Or it has to do with noise of some calibration method